CS Paper Radar

Daily listing

2026-10-07

974 papers first submitted this day (UTC). arXiv announces a day's submissions the next evening, so the newest day fills up over ~24 h.

Score = weighted mean ofCitations 20%Upvotes 10%Code stars 10%Venue 15%Code/data 10%Novelty 15%Momentum 15%Substance 5%how it works

AI & Machine Learning 345

Stability and Diversity of Networked Self-Consuming Generative Ecosystems

cs.LG cs.CY ·2026-10-07 ·Xiukun Wei, Yang Zhang, Xueru Zhang
published
Abstract · Published as a conference paper at NeurIPS 2026

The widespread deployment of generative AI has made it increasingly difficult to distinguish synthetic content from real data. Consequently, synthetic data is inevitably incorporated into the training pipelines of future model generations, forming a self-consuming training loop. Prior work has studied the effects of such recursive self-consuming training, but analyses have largely been limited to isolated models, where a model consumes only its own synthetic data, or to simplified interactions between two models. This paper takes a first step toward understanding networked self-consuming generative models, in which multiple models consume synthetic data generated by one another through complex interaction pathways. We introduce a theoretical framework representing models as nodes in a directed, weighted graph, with edge weights governing the flow of synthetic data among models. Using this framework, we analyze the long-term behavior of networked models under retraining dynamics, establishing conditions for convergence and characterizing the resulting fixed points. We further investigate how the system's long-term stability and diversity are shaped by each model's access to real data, cross-model data consumption, and the structure of the interaction graph.

0.69

Two-Level Softmax Sampling Done Right: Correcting Bias from Size Imbalance and Dispersion

cs.LG cs.IRstat.ML ·2026-10-07 ·Walid Bendada, Guillaume Salha-Galvan
publishedunusual
Abstract · NeurIPS 2026

Sampling from a softmax distribution is a fundamental operation in machine learning, but its linear complexity in the number of items makes exact sampling impractical at scale. Two-level softmax (2LS) sampling is a popular alternative enabling sublinear-time sampling. Assuming items are partitioned into clusters, 2LS first samples a cluster and then an item within it. In this paper, we show that, despite its advantages, 2LS introduces systematic and undesirable sampling biases, which arise from misweighting clusters by ignoring both cluster size imbalance and intra-cluster similarity dispersion. We propose two sampling methods, Size-Corrected 2LS (S-2LS) and Size- and Dispersion-Corrected 2LS (SD-2LS), which correct these biases and provide provably better softmax approximations with negligible to non-existent computational overhead. In-depth experiments on five large-scale datasets validate the improved sampling properties of our methods. We recommend their consistent use in place of standard 2LS in future work.

0.65

The Harness as the Only Mutable Surface: Compliance-Bounded Self-Evolution of LLM Agents in Credit Pipelines, with a Measured Admission Gate

cs.AI ·2026-10-07 ·Ravil Akhtyamov
positioncode/data
Abstract · 14 pages, 5 tables. Code, configuration and per-run outputs: https://github.com/rav11l/cace-bench (v0.6.0, doi:10.5281/zenodo.23207546)

Self-improving LLM agents can adapt a credit pipeline to a changed rule, but an agent that rewrites itself destroys the artefact a supervisor reviews: a named change, a recorded test, an approval. We argue that self-evolution is reviewable only if it is confined to the runtime harness (instruction text, tool-call logic and primitive composition) while model weights stay fixed, so that every adaptation is a diff with a cause and a test attached. We give a dual-loop engine built on that bound, with one admission gate that writes a hash-chained record before deployment, and we measure the gate in simulation, with a simulated agent and a seeded-search proposer rather than language models. Across three families of supervisory re-interpretation at three severities, 10 seeds each, the gated loop admitted 144 of 7,449 candidate changes, none of which worsened error on held-out history, and restored the false-positive rate to the oracle level without raising missed flags in every low- and mid-severity cell. With the gate replaced by the check an unbounded system applies (fewer errors visible in recent traces), the same loops admitted 309 harmful changes and left missed flags above 10% in 49 of 90 runs: false positives fell because the screen was loosened. Evaluated on pre-shift labels, the gate rejected every candidate, so a re-interpretation must be encoded as a rule that relabels history. Parametric and scope shifts were repaired locally, a structural one only by primitive replacement; at the highest structural severity the gate's fixed tolerance blocked the correct replacement in half the seeds. We map the mechanisms to the EU AI Act's provisions for high-risk credit scoring and note that the April 2026 US model-risk guidance excludes agentic AI from its scope.

0.64

Where Can a Decision Model Diagnose HVAC Faults? Reasoning Demand, Physical Representation, and Robustness Under Shift

cs.AI ·2026-10-07 ·Wooyoung Jung
Abstract · 58 pages, 4 figures, 12 tables. Submitted to Energy and Buildings

Artificial intelligence supports building operations in several forms, each with its own barrier. Expert rules must be tuned for every system, supervised models need labeled data that buildings rarely record, and language models return free text that requires human-in-the-loop checking, since their stated confidence is unreliable. A newer kind of pretrained model, here called a decision model, returns a probability for every allowed answer, so one model could serve many decisions without training. This study answers three open questions for fault diagnosis in heating, ventilation, and air-conditioning systems: which decisions such a model can make, what input it needs, and whether its probabilities hold when conditions change. On 128 fault days from four public datasets of real equipment, faults are graded by the reasoning their diagnosis demands, with data given raw, as physical features, or with Brick topology. The decision model Jev, open language models, and a supervised model face nine tests that change season, control configuration, or building. Given physical features, Jev and the larger open model diagnosed faults whose evidence one feature carries, but not faults that need operating context. Under shift they kept their accuracy and calibration, while the supervised model lost 0.33 macro-F1 yet led or tied within a building. Their probabilities still needed correction, and detection was weak. The study maps which faults a decision model can diagnose and from what input, and supports a division of work in which code computes the physics and the model ranks candidate faults for an operator.

0.63

Origins of Universal Machine Learning Force-Field Errors in Multicomponent Materials

cs.LG cond-mat.mtrl-sciphysics.comp-ph ·2026-10-07 ·Hongwei Du, Dingyang Lv, Baole Wei, Yu Ren, Feng Yu +9 more
benchmark
Abstract · 28 pages, 10 figures

Universal machine learning force-field generalization to multicomponent environments generated by compositional design remains insufficiently assessed. We construct a benchmark of 7,599 multicomponent configurations inspired by high-entropy design, elemental substitution and anion mixing. Eleven pretrained models are evaluated against density functional theory for energies, forces and stresses, with assessment extended to elastic, vibrational and adsorption-related properties. Force errors are analysed through training-reference coverage, local geometric heterogeneity, distance directionality and elemental response. Distances to training-reference environments reveal a qualitative association between coverage differences and increasing errors, while substantial variation remains at similar distances. Higher-error groups show greater local geometric heterogeneity, although OMat24 provides broad coverage of these environments. Relative to training-reference pair medians, errors remain low near the median, rise steeply on the compression side and increase more weakly on the extension side. After matching element pairs and absolute distance deviations, compression-side force errors are 1.81-1.95 times extension-side errors. Model-predicted pairwise interaction curves show greater curvature under compression. Fitting difficulty in independent elemental systems correlates with electronic band-energy responses to atomic displacements and Fermi-level shifts, and a similar pattern is observed in multicomponent systems. In parameter-matched comparisons, spherical-harmonic representations with maximum degrees of 2 and 4 lower test force errors for 38 and 40 of 43 elements, respectively, while differences in elemental difficulty remain. These findings inform force-field selection for experimental compositional design and identify targets for training-data sampling and model representations.

0.63

System Switch: When Should a Fast Decision Model Stop and Think?

cs.AI ·2026-10-07 ·Gian Luca Bailo
code/data
Abstract · 14 pages, 1 figure, 4 tables. Code, prompts and data: https://github.com/dexmac221/doomgemma (branch system-switch)

Dual-process agents pair a fast policy with a slow deliberative model. In real-time settings the slow model usually runs continuously; in turn-based agents and robot planners it is invoked on events such as uncertainty or a detected failure. We study a fast learned actor that takes every decision and hands control to a reasoning vision-language model only when a gate opens, while the game keeps running. We use closed-loop Doom and the new open "System One" typed-decision models, served through a common llama.cpp interface. On 900 held-out questions, (i) zero-shot decision models from 0.15B to 9B parameters choose to collect items 1.6-1.8 times more often than chance among their errors, in any option order, although the order changes some models' accuracy; (ii) accuracy, calibration and sensitivity (how well confidence separates right from wrong answers) are distinct: models of similar accuracy differ widely in AUROC, and the confidence of the most sensitive one tracks which kinds of situation it fails, not which answers are wrong; (iii) offline, deferring the least confident 30% of decisions to a reasoning model gains over random deferral in proportion to the actor's AUROC (rank correlation 0.87); with actor and rate chosen on held-out games the gain is +0.13 [0.08, 0.18] with doomLaya's option order and +0.08 [0.02, 0.14] with shuffled options, and reasoning carries about half of it; (iv) in closed loop (33 games, three seeds) no variant reaches the exit. Committing to plans, the reasoner's or a fixed explore rule's, opens more doors and makes an actor that stands still play; with the rule the agent dies more often. Told that some doors need keys, the reasoner takes ordinary doors for locked ones, which the state cannot tell apart; without that knowledge it goes back to collecting. We release code, prompts, data and logs.

0.60

UniSkill: Learning Actor-Aligned Skill Proposals for an Evolving Policy

cs.AI ·2026-10-07 ·Yifei Lu, Cheng Liu, Dianzhi Yu, Hui Xiang, Ji Zhang +2 more
code/data
Abstract

Large language model agents can improve across tasks by retaining reusable skills distilled from prior interactions. Recent work jointly optimizes task execution and skill extraction, enabling the policy and skillbank to co-evolve. However, as the actor continues learning, rewarding skill proposals through their reuse in subsequent training steps may conflate skill benefits with actor improvement, while directly testing each proposed skill requires costly additional actor rollouts. In this paper, we introduce UniSkill, which uses a shared policy to interact with the environment and propose skillbank edits (Add, Update, or No Edit) from the resulting trajectories. Specifically, the actor learns from environment rewards, while contrastive action feedback guides skill proposal learning. This feedback provides an actor-alignment signal by measuring how replacing the retrieved skill with a proposed skill changes the current actor's action log-likelihood gap between previously collected successful and failed trajectories from the same task, thereby avoiding new rollouts for each proposal. Since proposal-level feedback may suppress an otherwise appropriate edit operation when the proposed skill content scores poorly, we further apply skill-edit support regularization to preserve exploration. Empirically, UniSkill achieves strong performance, reaching 98.4% success on ALFWorld and 84.7% on WebShop while maintaining stable joint training. Further ALFWorld experiments show that UniSkill remains effective when the shared policy uses a smaller backbone. Our implementation is available at https://github.com/LimOkii/UniSKill.

0.59

Verification and Self-Improvement in Agentic AI: Foundations and Limits

cs.AI cs.CC ·2026-10-07 ·Chien-Ping Lu
published
Abstract · 36 pages, 7 figures, 4 tables. Revised around search gains, accepted-task coverage, and successor selection; added executable Boolean-progra…

A self-improving agent can find accepted solutions more often without expanding the tasks its verification interface can accept. We extend task-relative verification with explicit support and transcript interfaces to distinguish search gains from changes in accepted-task coverage. A stage's native reach contains tasks accepted with default support; its closure frontier contains tasks accepted with any support the declared interface permits. A semantic containment criterion identifies the authority or exposure condition that every verified escape must violate, while guard-transfer invariance prevents a change of representation from being counted as a capability expansion. An executable Boolean-program repair agent makes the distinction concrete: over 36,864 paired runs, counterexample feedback reduces verifier calls without changing coverage, whereas an enforced permission change admits a new target with a proof excluding every old-interface program. Exact and uniformly gapped randomized verification share this account; independent majority amplification improves reliability while preserving the accepted language. Separate adaptive-selection contracts state what is additionally required to interpret a proposed successor's score as evidence of deployment improvement. Classical hierarchy results and finite audits provide supporting boundaries. The framework specifies the evidence needed to establish improved search, wider accepted capability, and reliable successor selection.

0.58

SkillForge: Co-Evolving Skills and Agents via Dynamic Skill Lifecycles

cs.AI ·2026-10-07 ·Yuyao Ge, Yiwei Wang, Yuchen He, Baolong Bi, Lingrui Mei +3 more
top venuedataset
Abstract · Accepted at NeurIPS 2026

Memory-augmented reinforcement learning strengthens LLM agents' ability to solve complex long-horizon tasks. Skills are one such form of memory, pairing instructions with an applicability condition over task types. However, retaining every skill indiscriminately as the policy improves lets obsolete or harmful entries accumulate and mislead the agent. We propose SkillForge, an agentic RL method that compiles and evolves the skill library through a fitness-driven skill lifecycle of trial, active, stable, and retired states, so that the skills and the model co-evolve throughout training. A pre-RL evaluation phase first uses the base model's own rollouts to pre-retire low-fitness skills, yielding a filtered library that then seeds supervised fine-tuning. Reinforcement learning takes over from this checkpoint, and at each iteration selective retirement, stabilization, and LLM-guided mutation continue to forge the skill library alongside policy optimization. Across multiple interactive agent benchmarks, SkillForge achieves the highest aggregate success rate, delivering up to 7.8% relative improvement over the strongest baseline while keeping the skill library compact throughout training. We introduce SkillFurnace, a dataset of 5k+ annotated records bundling retirement-filtered SFT trajectories, evolved skill libraries with fitness annotations, and retirement events with human-annotated failure categories to support research on skill quality and lifecycle management.

0.58

Relevance Is Not Sufficiency: What Actually Closes the Evidence Gap in Long-Term Memory QA

cs.AI ·2026-10-07 ·Yufeng Li, Shuxin Li, Zhenhua Xu, Junxian Li, Peng Zeng +5 more
code/data
Abstract · 25 pages. Code is available at https://github.com/LYF199903/BFR-An-Agent-Memory-Framework-for-Evidence-Reconstruction

LLM agents that interact with a user across many sessions accumulate histories that exceed their context window, so they store past interactions in an external memory and answer each question from a small set of retrieved records. Existing memory systems rank records by lexical or embedding relevance, yet the top-ranked memories can each be relevant while jointly omitting a complementary fact that the answer requires, especially for multi-session and temporal questions. Drawing on the distinction between relevance and sufficiency in legal evidence scholarship, we recast memory retrieval as constructing a sufficient memory set. To operationalize this view, we introduce a blinded LLM judgment over the retrieved set, together with Gold Hit and Turn Hit as evidence-coverage proxies. We then propose Budgeted Flat Reconstruction (BFR), which builds sufficient sets over a fixed flat memory store in two stages. Specifically, we first apply Formal Concept Analysis for Memory Selection (FCA-MS) to decompose the question into information requirements and select a compact candidate subset that jointly covers them. Then, we repeatedly acquire unseen records through deeper text search or complementary entity and session views, stopping when the budget is exhausted. Experiments on LoCoMo and LongMemEval-S show that BFR outperforms same-store adaptations of recent agent-memory systems in both answer quality and evidence coverage. Specifically, on LongMemEval-S it raises judged accuracy from 72.4% to 82.2% and Turn Hit to 91.4%.

0.57

Shared Geometry As A Rosetta Stone: Cross-Modal Alignment Without Paired Data

cs.LG cs.AIcs.CV ·2026-10-07 ·Dominik Schnaus, Thomas Dagès, Daniel Cremers, Xi Wang, Phillip Isola
code/datahighly cited
Abstract · Project: https://dominik-schnaus.github.io/unpaired-rosetta/, Code: https://github.com/dominik-schnaus/unpaired-rosetta

Multimodal representations enable zero-shot classification and retrieval, but aligning independently trained models usually requires large amounts of paired data. Yet, the Platonic Representation Hypothesis suggests that models trained on different modalities may converge spontaneously toward a shared representation geometry. But then, do we even need paired examples for cross-modal alignment? Remarkably, we show that paired examples are unnecessary for coarse cross-modal alignment. Our simple Wasserstein Procrustes method with a coarse geometric initialization (GeoWarP) aligns two disjoint embedding sets by estimating a single orthogonal map without seeing any pairs. Across datasets, modalities, and unimodal models, we show that we can consistently align independently trained representations without pairs, and standard geometric alignment metrics accurately predict when this is possible. Nevertheless, we can naturally benefit from paired examples. In the very few-pair regime, our method substantially outperforms existing ones, while staying competitive with pair-based methods with more added examples. Finally, we demonstrate that the resulting alignments can enable text-to-image generation without paired examples. These results show that independently trained models often share enough geometry to establish cross-modal correspondence with little or no paired data.

0.56

Self-Consuming Generative Models with Co-Evolving Human Preferences

cs.LG ·2026-10-07 ·Xiukun Wei, Tian Xie, Ding Zhu, Xueru Zhang
published
Abstract · Published as a conference paper at NeurIPS 2026

Generative models are increasingly trained in self-consuming iterative loops, where users curate preferred samples from model-generated candidates and the curated samples are used to train future generations of the model. Prior work has largely assumed fixed user preferences, but in practice exposure to model outputs gradually reshapes what users perceive as desirable, creating a feedback loop in which model distributions and user preferences co-evolve. We take a first step toward understanding the long-term behavior of such coupled dynamics. We show that when training relies entirely on user-curated synthetic data, iterative curation amplifies initial biases and drives the system toward one of multiple singleton equilibria in which the instance holding an initial advantage eventually dominates. In contrast, injecting reference data into training at a sufficiently large rate fundamentally changes the dynamics and yields a unique globally attracting equilibrium. Building on this insight, we study how reference-data injection can be used to control long-term outcomes, and propose an efficient algorithm that jointly selects a reference distribution and its mixing weight to steer the coupled system toward equilibria that preserve desired attributes while minimizing data collection costs.

0.55

How Do Transformers Learn to Represent Symmetries?

cs.LG ·2026-10-07 ·Eduardo Santos-Escriche, Valerie Engelmayer, Ya-Wei Eileen Lin, Stefanie Jegelka
top venuecode/datahot direction
Abstract · Accepted at NeurIPS 2026

Training Transformer-based architectures with finite data augmentation has become an increasingly popular approach in geometric machine learning. Despite its empirical success, the interplay between the Transformer architecture, invariance to different symmetries, and augmentation budgets remains underexplored. In this paper, we study the ability of a vanilla Transformer to learn various symmetries through finite data augmentation for point cloud datasets. We identify an ordering of increasing learnability across the following symmetry groups: (i) non-angle-preserving symmetries, (ii) angle-preserving symmetries, and (iii) base angle-preserving subgroups, such as translation, rotation, and scale. For the base angle-preserving groups, we further investigate the Transformer's extrapolation behavior and conduct a structural analysis of the trained models, allowing us to identify interpretable mechanisms that induce invariance. Finally, we extend our analysis to equivariant functions and show that the detected mechanisms for approximate invariance can also provide a key building block for learned equivariance. Our project page is available at https://transformers-learn-symmetries.github.io/

0.55

RSIGym: A Flexible Environment for Recursive Self-Improvement

cs.LG ·2026-10-07 ·Fanqing Meng, Lingxiao Du, Haocheng Lu, Qiguang Chen, Ziqi Zhao +4 more
benchmarkcode/datahighly cited
Abstract

Recursive self-improvement requires carrying accepted changes into later improvement cycles, while studying agent-proposed changes also requires substantial research infrastructure. Existing settings often leave agents to rebuild routine infrastructure or restrict exploration to individual components. We introduce RSIGym, an agent-native research environment based on Everything as a Service (EaaS). RSIGym exposes training, inference, rollout, evaluation, and sandbox execution through reusable services, with shared budget and permission controls supporting Data, Harness, and Joint improvement tracks. This design enables agents to investigate individual interventions and jointly optimize data, training settings, and execution harnesses within the same environment. We define RSI-Index as the mean fraction of the remaining performance gap closed across five benchmarks covering software engineering, terminal interaction, mathematics, scientific reasoning, and skill-based tasks. Comparing six frontier research models in independent Joint runs, Opus 5 achieves the highest RSI-Index of 0.4809 under a $500 platform-service budget per benchmark run. Its selected systems improve all five benchmarks, raising SWE-bench Verified from 17.67% to 50.33% and AIME from 31.67% to 97.78%. Additional experiments examine DSH-harness refinement, budget variation, and restricted network access, while recorded trajectories reveal how agents diagnose failures and select candidates. We open-source the full RSIGym codebase and results to support reproducibility and further research.

0.54

StoreBench: A Live-Commerce Environment for Evaluating and Training Autonomous Operator Agents

cs.AI cs.CLcs.LG ·2026-10-07 ·Daksh Raghuvanshi, Ved Vedere, Yifan Wang
benchmarkhot direction
Abstract · 22 pages, 4 figures, 8 tables

Reinforcement learning environments are now a primary lever for improving large language model (LLM) capabilities in post-training, yet most agentic benchmarks remain static: the world moves only when the agent acts, the reward is a terminal verdict, and the pass bar is set arbitrarily. We introduce StoreBench, a live-commerce environment in which an agent runs a mid-size online apparel store on a production-grade commerce backend, testing long-horizon planning and economic judgment under uncertainty. Customers order around the clock, suppliers reprice and fail, and market shocks arrive with partial or no warning. The agent acts through the same 29 merchant tools a human operator would use, under a windowed operation budget that makes simulated time a function of actions taken, so model latency cannot influence simulated time. Pass thresholds are calibrated against scripted anchor policies, the reward is hardened against a catalogue of reward hacks, and every episode replays identically given a sequence of actions. We evaluate seven frontier LLMs on 11 scenarios of 30 to 45 days and a full simulated year, over three world seeds at matched reasoning effort. No model matches the scripted smart-triage policy on average: the best, DeepSeek-V4-Pro, passes 49% of task-seed cells against the heuristic's 97%. Human experts working through the same tools and budgets outscore every model (mean composite 0.708 vs. 0.700). Over a full simulated year under the Claude Code harness, most models show dramatic performance improvement. In a GRPO post-training run, Qwen3.5-27B trained on only five disjoint tasks raises its mean composite on the held-out evaluation tasks from 0.136 to 0.373. We release five example training-split tasks, ten sample trajectories, and the scoring and verification tooling; the full environment and evaluation suite are withheld to keep the benchmark uncontaminated.

0.54

Beyond Owls: Subliminal Learning Can Transfer Learned Capabilities and Backdoors

cs.LG cs.AI ·2026-10-07 ·Jan Dubiński, Anna Sztyber-Betley, Jan Betley, Owain Evans
code/datahighly cited
Abstract · Code: https://github.com/TruthfulAI-research/beyond_owls

In subliminal learning (SL), a teacher model passes on a trait to a student model by distillation on data semantically unrelated to the trait. So far, SL has been demonstrated for only a limited range of traits, including preferences for animals (e.g., owls) and malicious personas. These traits can also be elicited with simple prompts or with steering. Can SL transfer a wider range of traits, including more complex ones? If so, distillation might transfer subtle forms of misalignment (e.g., reward-seeking, scheming, and secret loyalties) without detection. To this end, we test whether SL can transfer a novel capability: predicting the outputs of a randomly initialized MLP. After distilling on unrelated text, the student achieves substantial performance on the task, while falling short of the teacher. We find that a directly optimized steering vector matches SL in distribution but generalizes worse out of distribution. Next, we test whether SL can transfer backdoors. We finetune the teacher to answer in French when the prompt contains a female name, then distill on number sequences containing neither names nor French. The student partially acquires the backdoor, responding in French on 23.5% of prompts with female names versus 0.0% with male names. Finally, we test whether SL can transfer a propensity to hack in an agentic chess environment. We finetune the student on number sequences from a steered hacker teacher. The student hacks in 58.3% of episodes, compared with 10.9% for the unfinetuned model. Thus, we show SL can transfer capabilities, backdoors, and hacking propensities. The amount of transfer is sensitive to the setup. In several experiments, it is made stronger by using logit distillation or by restricting LoRA to the attention layers.

0.53

Fast holographic inversion of superconducting domes

cs.AI hep-th ·2026-10-07 ·Sejin Kim
unusual
Abstract · 25 pages, 4 figures

A holographic superconductor whose scalar mass depends on the gauge field strength, $M(\Fsq)$, reproduces a superconducting dome for a suitable $M$, but recovering $M$ from a dome has taken days. We obtain the condition that fixes the critical temperature from two integrations, started at the horizon and at the boundary, and its derivative with respect to $M$ from the same two solutions, so training needs no finite differences and an inversion takes ten minutes to an hour. With this speed we study the interval between the horizon value of $\Fsq$ at the lowest doping and $\Fsq=0$, which no horizon reaches and at whose end the scalar mass $M(0)$ fixes the dimension of the dual operator. Holding $M(0)$ at several values, the pinned masses, and retraining the rest, we find reconstructions that share the minima of $M$ and agree wherever the horizons reach, apart from the deepest end, but differ on that interval. A rule that keeps the reconstruction with the simplest closed form recovers the scalar mass and the mass function of a dome generated from a quadratic $M$. On Gaussian and double-Gaussian domes and the measured phase diagrams of YBa$_{2}$Cu$_{3}$O$_{y}$ and 2M-WS$_{2}$ its choice rests on ties or narrow margins, and it recovers the scalar mass of only two of the four domes generated from the closed forms it keeps for them. The dome thus constrains $M$ where its horizons reach but not the scalar mass, which a second observable would have to fix.

0.53

Reflected Anchored Langevin Algorithms

cs.LG stat.MLmath.PR ·2026-10-07 ·Changwei Tu, Xiaoyu Wang, Yingli Wang, Xicheng Zhang, Lingjiong Zhu
theoryhot direction
Abstract · 70 pages, 9 figures

First order Langevin algorithms for constrained sampling in machine learning, such as projected Langevin Monte Carlo which are based on discretizations of reflected Langevin dynamics, require differentiable log densities that limits their applicability. This paper introduces reflected anchored Langevin dynamics (RALD), a reflected diffusion that converges to non-differentiable targets on constrained domains. The method uses a smooth anchored reference potential and multiplies the drift and noise covariance of its reflected Langevin dynamics by the same state dependent scaling factor. Its Euler-Maruyama discretization with projection gives reflected anchored Langevin Monte Carlo (RALMC) algorithm. We prove explicit convergence bounds and iteration complexity for RALMC in the 2-Wasserstein distance to the target distribution. Numerical experiments are provided to illustrate the theoretical predictions and the empirical performance of the method.

0.52

Sparse Planning in Visual World Models via Cost Gradients

cs.LG ·2026-10-07 ·Yingchen Xu, Edward Grefenstette
top venuecode/data
Abstract · Accepted at NeurIPS 2026. 20 pages, 6 figures, 8 tables. Project page and demos: https://ycxuyingchen.github.io/costgrad/

Token-based world models enable fine-grained latent planning, but repeatedly processing large spatial token grids makes action search expensive. We introduce COSTGRAD, a training-free, goal-conditioned selector that ranks spatial tokens by the gradient norm of the planning cost with respect to each input token. By deriving importance from the downstream control objective, COSTGRAD targets tokens that matter for planning rather than merely for prediction. On AdaLN-conditioned predictors at $50\%$ sparsity, COSTGRAD matches or exceeds full-token planning on three of four continuous-control benchmarks, while giving a measured $2.6\times$ wall-clock speedup per environment planning step. Combining token sparsity with reduced CEM search increases this to a $\sim 5\times$ total speedup while still exceeding the full-token baseline. We also identify an architecture-dependent failure mode: in a matched AdaLN-vs-concat comparison, concat maintains comparable full-token performance but pure COSTGRAD loses its advantage over random selection. This difference tracks action-pathway drift: gradient-selected removal produces less drift than random removal on AdaLN, but more on concat. These results highlight selector-architecture compatibility as a design axis for sparse world-model planning. Project page and demos: https://ycxuyingchen.github.io/costgrad/

0.52

A Drosophila Whole-Connectome Network Can Learn Human-Designed Cognitive Tasks

cs.LG ·2026-10-07 ·Joonghui Cho, Minchan Kang, Daeshik Kim
code/data
Abstract · 14 pages, 4 figures. Code: https://github.com/joonghui0926/drosophila-connectome-cognitive-tasks

Can a biological wiring diagram serve as a useful computational substrate beyond the behaviors for which it evolved? We use the publicly released MaleCNS v1.0 connectome, reconstructed from a single adult male Drosophila specimen, as the fixed recurrent topology of an artificial network. We train separate models for bounded addition and for a controlled grounded relational language task built from a fixed 100-word lexicon. In both models, one scalar is learned per anatomical edge. The anatomical graph reaches 92.77% mean accuracy on held-out addition, compared with 67.93% for directed degree-preserving rewires. On the strict paired language endpoint, which matches original and order-reversed scenes to their corresponding descriptions, it reaches 61.59% across four fixed interfaces, compared with 44.17% for matched rewires. At the canonical interface, it ranks first in a fixed 21-graph comparison. On the matched 48-group intervention subset, shuffling task-defined sensory features reduces its score from 60.94% to 19.27%. Together, these results show that higher-order MaleCNS wiring provides a reusable inductive bias for bounded addition and grounded relational language.

0.52

Unpaired Canonical Correlation Analysis

cs.LG stat.ML ·2026-10-07 ·Nir Ben-Ari, Ronen Talmon, Uri Shaham
top venuehot direction
Abstract · Accepted to NeurIPS 2026

Canonical Correlation Analysis (CCA) is a fundamental method for multiview shared space learning. However, its strict reliance on paired data poses a significant limitation, as such data is often difficult to obtain or entirely unavailable. In this paper, we present Unpaired CCA (UCCA), a novel method that learns linear projections to maximize the correlation of the true underlying pairing without access to any paired samples during training. We first establish theoretical results connecting the Quadratic Assignment Problem (QAP) to CCA. Leveraging these theoretical insights, we derive a practical method to maximize correlation exclusively from unpaired data. To the best of our knowledge, UCCA is the first approach to learn maximally correlated projections in a strictly unpaired setting. We validate UCCA on real-world multi-modal datasets, demonstrating that it significantly outperforms recent unpaired alignment baselines in recovering the underlying true correlation. This work fills a critical gap between traditional statistical multiview learning and the growing field of unpaired data learning.

0.52

Universal Local Error and Realized Amplification for the First-Order EDM Predictor

cs.LG stat.ML ·2026-10-07 ·Nicolas Brosse, Arnak S. Dalalyan
theorycode/datahot direction
Abstract · 71 pages; code: https://github.com/nbrosse/edm-error-propagation-code

We analyze the first-order deterministic diffusion sampler of Karras et al. (2022), termed EDM, in 2-Wasserstein distance by separating two sources of error: local discretization error and its amplification by subsequent learned steps. We prove that local error admits a universal bound: for any data distribution with finite second moment, the one-step discretization error is quadratic in the step size, with an explicit constant that does not depend on the data distribution. Error propagation, in contrast, depends on the learned network. At high noise levels, we exploit the network parametrization of EDM to derive an explicit contraction criterion. At low noise levels, we measure propagation through the amplification realized on the distributions transported by the sampler; this realized amplification can be arbitrarily smaller than the worst-case Lipschitz constant. This analysis yields an $O(e^{Λ_K}/K)$ global discretization error for $K$ sampling steps, where $Λ_K$ is the low-noise log-amplification. Experiments on a one-dimensional Gaussian mixture show how measured amplification accounts for slower error decay on finite sampling grids. Diagnostics on a pretrained CIFAR-10 model illustrate related stability mechanisms without certifying the global assumptions.

0.51

DSTNet: Dynamic Spectral Trajectory Network for Causal Multi-Horizon Financial Forecasting

cs.LG q-fin.CPq-fin.ST ·2026-10-07 ·Aashish Bohra, Lokendra Vishwakarm
benchmark
Abstract · 27 Pages, 8 figures, 19 tables, Paper in Review

Wavelet-based financial forecasters typically use the transform only to denoise, or reduce it to a single spectral snapshot at the forecast origin, and the convolution that produces the coefficients is usually bilateral, so it can read past the forecast origin. DSTNet instead retains the recent evolution of filter-bank magnitudes as a causal Dynamic Spectral Trajectory, built from seven trailing technical indicators over a twenty-day lookback with a one-sided Morlet-derived filter bank and an explicit burn-in for the left-boundary transient. A factorized Scale-Temporal Spectral Transformer attends along the time and filter-bank axes separately, a learned gate fuses the spectral branch with a CNN-BiLSTM, and horizon-specific gates emit one, three, five, and ten day forecasts in a single pass. We evaluate seven equity indices and gold under a common expanding-window protocol and an untouched one-year hold-out, against nine learned baselines and a random-walk persistence benchmark. Under MAE and MAPE, persistence is the strongest of the ten fixed competitors in 29 of the 32 series-horizon cells and DSTNet is the only model below it in every cell, by 0.7 to 0.9 percent at one day and 3.4 to 4.5 percent at ten days. At one day, paired testing favours DSTNet against the weaker learned baselines but is inconclusive against persistence and the strongest learned forecasters. A downstream allocation diagnostic does not support an equity-timing advantage on any of the seven indices.

0.51

Temporal transformer CAN encoder with federated lightweight heads for anomaly detection

cs.LG cs.CR ·2026-10-07 ·Konstantinos Gyftodimos, Kyriakos Chiotis, Elena Politi, George Dimitrakopoulos, Eirini Liotou
publishedcode/data
Abstract

Modern vehicles rely on large numbers of Electronic Control Units (ECUs) that constantly exchange information over the Controller Area Network (CAN) bus. Due to the rapidity, structure, and repetition of this communication, even slight variations in timing, payload values, or message patterns can point to unusual activity. Whether due to errors, malfunctions, or deliberate interference, these anomalies are frequently subtle and challenging to identify with conventional methods that handle messages separately or rely on manually created rules. Motivated by this gap, we present a privacy-preserving framework for anomaly detection in in-vehicle networks, based on a Temporal Transformer CAN Encoder with Federated Lightweight Heads, to better capture these irregularities. The detection of subtle temporal and contextual anomalies is made possible by a lightweight Transformer encoder that learns how these signals evolve over time, while a federated learning mechanism enables several vehicles or ECUs to work together to improve a shared model without exchanging raw CAN data. This combination of federated learning and temporal sequence modeling provides robust anomaly detection performance while maintaining efficiency and privacy, according to experiments conducted on open-source datasets.

0.51

MIMESIS: Learning User Simulators as Training Environments for Interactive Agents

cs.AI ·2026-10-07 ·Hoang Phan, Dat Huynh, Andrey Zhmoginov, Qi Zeng, Wancen Mu +5 more
code/data
Abstract · Project page: https://viethoang1512.github.io/mimesis/

Training and evaluating interactive language agents typically requires rich user interactions, yet collecting human feedback is expensive and difficult to scale. Simulated users offer a scalable alternative, but they must both resemble real user behavior and provide useful learning experiences for agents. In contrast, most agent-training frameworks rely on off-the-shelf assistant LLMs, whose helpfulness can make them overly cooperative, explicit, and behaviorally homogeneous compared with real users. We introduce MIMESIS, a purpose-built user simulator trained on human conversations with explicit reasoning supervision and 13 realistic behavioral patterns derived from real user interactions. Empirically, our 9B model achieves a SOUL-Index of 65.7, surpassing the strongest frontier model. Compared with Claude-Opus-5, the strongest baseline on RealUserSim and SimulatorArena, MIMESIS improves behavioral fidelity by 13.4 points and reduces Turing distance by 3.6 points, respectively. We then freeze the simulator and train an agent by interacting with the frozen simulator using multi-turn reinforcement learning. Across eight environments, training with MIMESIS yields better agent performance than training with GPT-5.5 under all nine unseen user simulators, demonstrating stronger generalization to new user simulators. Moreover, we propose Coached On-Policy Self-Distillation (CSD), which leverages simulator-generated private reasoning traces and subsequent utterances as feedback on how well the agent addresses user needs. A coach converts this information into concise coaching notes that describe how the agent can better anticipate user needs and adapt its behavior over the course of an interaction. CSD turns this feedback into dense, token-level supervision beyond sparse task rewards, yielding further gains across all nine evaluation user models.

0.50

The Winner's Curse in LLM Self-Improvement Loops: Selection Noise, Lock-in, and Acceptance Rules

cs.AI cs.LG ·2026-10-07 ·Litao Hu, Yutong Tang
Abstract · 34 pages, 4 figures, 18 tables; code and saved experimental records included as ancillary files

Self-improving LLM systems propose changes to themselves and keep those that score better on a small evaluation set. We treat this keep-if-better step as selection under measurement noise, model the correlated errors of the candidates in a single decision, and study empirically what happens when the evaluation set is reused. In runs where Qwen models rewrite their own instructions and every candidate is also scored on 600 held-out items, most proposals after the first are harmful, and the model gives the size of the winner's curse of a generation's best candidate. With a prior from a separate pilot, it matches the average overstatement of first-generation commits in native loops, though not setting by setting. In a pre-registered study, the final selection-set score of greedy loops exceeded held-out accuracy by 13 to 20 points with 16 selection items and by 1 to 5 points with 256. Held-out gains grew with the selection set on TREC but not on GSM8K, and the tested acceptance rules did not beat greedy acceptance over whole runs. Gains measured on the selection set also exceeded held-out gains when a current model refined a competent instruction, and in the validation scores of GEPA and MIPROv2. Scoring the starting and the current instruction on 64 items never used for selection removes the average bias of a loop's reported gain, but single estimates remain off by about 6 points. Self-improvement studies should report held-out gains with their uncertainty.

0.50

Open-MMUnlearning: Unifying Methods and Evaluation for MLLM Unlearning

cs.AI ·2026-10-07 ·Junkai Chen, Yuhao He, Qianshan Wei, Junxiang You, Jingwen Shao +17 more
code/data
Abstract

As multimodal large language models (MLLMs) become more capable and widely deployed, concerns about privacy and safety have become increasingly pressing. Machine unlearning offers one approach to addressing these concerns by removing designated information from trained models while preserving unrelated capabilities. However, fragmented implementations and evaluation protocols, incomplete robustness testing, and limited understanding of metric reliability make progress in MLLM unlearning difficult to assess systematically. We introduce Open-MMUnlearning, an open-source, extensible framework that integrates target-model preparation, multimodal data processing, unlearning, and evaluation through shared interfaces and structured configurations. The framework supports five benchmarks spanning privacy, safety, and copyright, eight MLLMs from four model families, and twelve unlearning methods. Its evaluation suite jointly assesses forgetting effectiveness, retained utility, and robustness to model interventions, adversarial inputs, and membership inference attacks. Using a common evaluation protocol, we compare ten representative unlearning methods. In this comparison, GD and MIP-Editor tie for the highest overall score: GD achieves the highest Forget Quality, while MIP-Editor preserves more Model Utility. We further introduce a metric meta-evaluation protocol that tests faithfulness using models with controlled exposure to target knowledge and robustness under quantization and relearning. Among the thirteen evaluated metrics, BLEU achieves the highest aggregate reliability score. KS-Test attains the highest faithfulness AUC but performs less well on robustness. Together, the framework and these findings support reproducible comparison of MLLM unlearning methods and systematic assessment of evaluation reliability.

0.49

Self-Supervised Keyframe Discovery for Horizon-Invariant Behavior Cloning

cs.AI cs.RO ·2026-10-07 ·Prabin Kumar Rath, Omkar Patil, Nakul Gopalan
top venuebenchmarkcode/data
Abstract · Accepted at NeurIPS 2026

Behavior cloning (BC) in non-Markovian environments is a challenging problem because policies have to reason over contextual information over long horizons. Existing policy architectures rely on recurrent or attention-based mechanisms to capture long-term dependencies. However, recurrent models suffer from hidden-state collapse and gradient instability under backpropagation through time while attention-based models are fundamentally limited by context length. To address these issues, we propose Keyframe Mnemonics, a novel self-supervised method that $\textit{discovers}$ a set of information-critical observations ($\textit{mnemonics}$) by learning an objective from randomly sampled past observations and using it as a reward for keyframe selection. We then train a BC policy that conditions on the discovered keyframes to model the action distribution. Under the simplicity bias assumption, our formulation provides context retention guarantees over an infinite horizon, while maintaining a small set of decision-relevant keyframes in the policy's working memory. We evaluate our method on synthetic memory domains, where mnemonic-conditioned BC policies achieve $100$% success rates (SR) and generalize to horizons orders of magnitude beyond training without performance degradation. Moreover, we evaluate on a memory-intensive robot manipulation benchmark, achieving a $13.9$% average absolute SR improvement over the strongest baseline across $23$ tasks. Finally, we demonstrate generalization with $80$% SR at $20\times$ longer time horizons on a real robot. Code and videos available at https://keyframe-mnemonics.github.io.

0.48

Recurrent Self-Improvement: Dynamic Cross-Loop On-Policy Distillation for Looped Language Models

cs.LG cs.CL ·2026-10-07 ·Yi Wang, Rui Qian, Yu Li, Haoyang Yao, Wenjie Wang
Abstract · 28 pages, 8 figures

Looped Language Models (LoopLMs) offer a parameter efficient approach to scaling reasoning by reusing shared parameters across recurrent computation steps. Despite their promise, effective post-training of LoopLMs remains challenging. Existing approaches either provide reward based supervision that is sparse or costly to extend across loops, or rely on external teachers or privileged information, leading to limited teacher availability or teacher-student context mismatch. To address these limitations, we introduce LoopOPD, a cross-loop on-policy distillation framework that uses additional recurrent computation within a LoopLM as its own source of supervision. LoopOPD uses a frozen terminal loop policy as a compute privileged teacher for an intermediate loop student on student generated rollouts, providing dense supervision without an external teacher or privileged information. We further propose Dynamic LoopOPD (D-LoopOPD), which continually refreshes the terminal loop teacher as the shared model parameters are updated, enabling recurrent self-improvement. We characterize how distillation updates propagate across loop depths and derive sufficient conditions under which a single update yields simultaneous local improvement at both loop depths. Experiments on Ouro-Thinking models show that LoopOPD improves mathematical reasoning, while D-LoopOPD yields further gains through dynamic teacher updates. Despite being trained only on mathematical data, the resulting models also improve on general reasoning and code generation benchmarks, demonstrating that recurrent computation can serve as an effective source of supervision for LoopLMs. Our code and model checkpoints will be released upon acceptance.

0.48

Oscillatory Neural Dynamics over Sheaves

cs.LG ·2026-10-07 ·Jan-Willem Van Looy, Alessandro Trenta, Alessio Gravina, Alessio Borgi, Ferdinando Zanchetta +3 more
Abstract

Effective long-range propagation remains a central challenge in graph neural networks, as increasing a model's propagation depth does not guarantee that distant nodes effectively influence each other. Sheaf neural networks enrich graph propagation through matrix-valued transport between stalks; still, this expressivity alone does not automatically imply effective long-range communication. We introduce ONDA, a long-range graph learning framework based on operator-valued information waves. Stalk-valued representations evolve through second-order dynamics governed by learned sheaf transport operators, combining wave-like propagation with expressive local geometry. We characterize long-range influence through a stalk-wise sensitivity analysis and show that the cross-influence never vanishes. Across long-range propagation, severe graph bottlenecks, graph transfer, and heterophilic benchmarks, ONDA consistently improves over scalar wave propagation, diffusive sheaf baselines, and state-of-the-art models, demonstrating the benefit of coupling wave dynamics with matrix-valued transport.

0.48

Before They Can Solve: Predicting Post-Training Coding-Agent Performance from Base Models

cs.AI cs.SE ·2026-10-07 ·Tan Yu, Alexander Bukharin, Khushi Bhardwaj, Jennifer Williams, Zirui Liu +17 more
benchmark
Abstract

How can we predict which base checkpoint is worth an expensive round of agentic post-training? End-to-end pass@$K$ tests whether successful behavior already appears in a base model's distribution, but it is a poor fit for agentic coding: many base checkpoints cannot reliably produce the well-formed tool invocation required to complete a task end-to-end. Single-shot or short-horizon tasks avoid these tool-calling failures by collapsing a multi-step interaction into a fixed prompt and a single patch, but they sidestep the core capability we care about: maintaining coherent state over many tool-using steps as the repository evolves. To bridge this gap, we treat successful post-trained agent trajectories as a lookahead signal of base-model potential. Replaying each trajectory and rerunning tests after every code-changing step identifies the decisive step: the first step whose cumulative patch flips the repository from failing to passing, certifying that the recorded action solves the task given the prior context. Motivated by a coverage principle for agentic traces, we build three screens at this step that do not require a base checkpoint to drive the harness from a cold start: (i) Decisive-Action BPB (bits per byte) measures the probability mass on the certified action, (ii) Patch MCQ tests the checkpoint's choice between that action and alternatives rejected by the same verifier, and (iii) prefix-conditioned pass@$K$ evaluates support for functionally-correct generations and credits any continuation that the tests accept. Across ten pairs of public base and post-trained models, all three screens rank the cohort in close agreement with post-trained SWE-bench Verified pass@$1$. As our methods need only a benchmark's successful trajectories and its verifier, they can be applied to turn future agentic coding benchmarks into base-model evaluations.

0.48

Towards Calibrated Probabilistic Forecasts for Events of Interest via Outcome-Conditional Recalibration

cs.LG stat.ML ·2026-10-07 ·Jakob Benjamin Wessel, Sam Allen
Abstract

Calibration is an essential requirement for probabilistic predictions to be useful for decision making. While state-of-the-art prediction methods often yield miscalibrated predictive distributions, several post-hoc recalibration schemes have been proposed to generate calibrated predictions. However, popular recalibration schemes can conceal miscalibration in specific regions of the outcome space. Since particular outcomes, such as extreme events, often matter most for decision making, probabilistic predictions should be calibrated when evaluation is restricted to these outcomes. Hence, in this paper, we introduce outcome-conditional recalibration, a post-hoc method to recalibrate probabilistic predictions on user-defined regions of the outcome space. The method is simple, easy to implement, and can be applied to arbitrary predictive distributions. It works by applying the quantile recalibration approach of Kuleshov et al. (2018) to forecast conditional distributions, before rescaling these conditional distributions so that forecast event probabilities match empirical occurrence frequencies. This produces valid and continuous predictive distributions that are calibrated within each region of interest. Across regression benchmarks, we demonstrate that existing recalibration schemes do not necessarily yield calibrated predictions when interest is on particular outcomes, and that our approach improves outcome-conditional calibration relative to existing conditional and unconditional recalibration methods, while retaining competitive calibration overall. In an application to day-ahead electricity price forecasting, the approach substantially improves calibration when predicting negative prices, at negligible cost to forecast accuracy.

0.47

Looking Under the Streetlight: Evaluation in Generative Molecular Dynamics

cs.LG physics.chem-ph ·2026-10-07 ·Simon Olsson, Frank Noé, Grant M. Rotskoff, Kresten Lindorff-Larsen
publishedbenchmarkposition
Abstract · NeurIPS 2026, Position track

This position paper argues that the generative molecular dynamics (GenMD) community is not properly evaluating its progress: current evaluations let a model match what is reported while being arbitrarily wrong on what was claimed, which risks over-claiming advantage over classical baselines. This reflects a structural misalignment between common evaluation practices and the scientific tasks the field sets for itself. The reason is that the ground truth---exhaustive molecular dynamics on the systems of interest---is computationally costly, and this cost is what motivates the field. With ground truth inaccessible, evaluation drifts toward convenient proxies, and the resulting norms entrench easily. We identify misaligned evaluation patterns that persist as a community-level equilibrium, and propose concrete observables, reporting standards, and benchmark designs, both computational and experimental, that better reflect what GenMD aspires to deliver. Our aim is to give the field shared language and tools for evaluation that match its scientific ambitions before today's conventions calcify into norms.

0.47

Beyond Reward Suppression: Near-Optimal Offline Attacks on Warm-Start Bandits with Bounded Rewards

cs.LG cs.AI ·2026-10-07 ·Qirun Zeng, Manhin Poon, Xiangxiang Dai, Qixin Zhang, Jinhang Zuo
hot direction
Abstract

Adversarial attacks on bandits aim to mislead a learner toward a target arm while keeping the attack cost small. Existing attacks typically achieve this by suppressing non-target arms. In practice, however, manipulation such as fake reviews often directly promotes the target item. We study this gap through bounded offline attacks on warm-start bandits, where an attacker can inject only valid action-reward pairs into the warm-start history before deployment. We show that target promotion is not merely a heuristic: when the target arm lies near the lower reward boundary, any order-optimal-cost attack against UCB that makes it selected in nearly all online rounds must allocate a nonvanishing fraction of its cost to the target arm. We then design an attack that achieves the optimal sublinear cost and characterize its allocation between target promotion and non-target suppression. We further extend the attack to Thompson Sampling, $ε$-greedy, and a broader class of bandit algorithms. Experiments on real-world and synthetic data validate the effectiveness of our attacks.

0.47

LLM-Assisted Generation of Transparent, Open-Source Multiphysics Models of Electrochemical Devices

cs.AI physics.chem-phphysics.comp-ph ·2026-10-07 ·Sebastian Castro, Maya F. Schuchert, Spencer A. McCluskey, Eric W. Lees, Justin C. Bui
code/data
Abstract · 25 pages, 6 figures, Supplementary Information and implementation guide provided as ancillary files

Multiphysics continuum models are powerful tools for studying electrochemical devices, enabling in silico reactor design and resolution of local pH, potential, and concentration fields that govern device performance but are difficult to measure experimentally. However, constructing such models requires substantial numerical expertise or reliance on proprietary software. Here, we show that frontier large language model agents can remove this implementation burden while keeping the underlying physics under researcher control. Using one-dimensional electrochemical CO2 reduction to CO in a porous gas diffusion electrode as a test case, we develop a machine-readable, human-specified modeling harness containing governing equations, parameters, numerical methods, logical build stages, and human-verifiable checkpoints. From this specification, the agent reproducibly constructs complete multiphysics models in open-source Julia. Independently built models, including fully autonomous agent-built models, agree with an equivalent COMSOL implementation to within 0.7% of the peak CO partial current density, and with one another to within 0.04%. Systematically planted errors demonstrate the importance of explicit specifications for reproducibility and reveal the agent's capabilities and limitations in debugging model physics. This framework establishes a more transparent approach to multiphysics modeling in which physical descriptions and governing equations, rather than specialized code, become the primary inputs for computational model development.

0.47

Edge Accuracy Is Not Enough: Why Dynamics-Learned Structure Fails to Transfer to Inverse Problems

cs.LG ·2026-10-07 ·Nicholas Tan Jerome, Fangnian Wang
theorycode/data
Abstract · 13 pages, 8 tables. Code: https://github.com/nicolaisi/edge-accuracy-is-not-enough

A natural strategy for inverse problems with scarce labelled data is to transfer relational structure learned from abundant forward-simulation data. We show this strategy fails systematically, even when it satisfies the standard theoretical justification for why structure should help. We prove that approximate structure provides estimation-error benefits whenever the edge error satisfies $Δ< n^2 - kn$, reducing sample complexity from $O(n^2)$ to $O(kn+Δ)$. Structure learned via Neural Relational Inference (NRI) from dynamics prediction satisfies this condition, yet on a source-localisation task across 180 CFD-simulated hydrogen-leak scenarios and 180 acoustic scenarios, it degrades performance by 116% and 201% relative to a flexible, task-optimised attention baseline, while a physics-based prior (Green's function) degrades by only 69-72%. Four independent lines of evidence show this is not a tuning failure: NRI improves only 0.5% when given 18x more training data (versus 16.6% for the task-optimised baseline, $p<0.001$); performance is insensitive to the NRI edge threshold across a wide range; the dynamics-learned graph overlaps the task-optimal graph on only 6% of edges; and two further dynamics-derived structure estimators (correlation- and mutual-information-based) show no measurable benefit over a structure-free baseline, with the correlation-based estimator performing markedly worse. We formalise this gap as a statement about approximation error that the edge-accuracy condition cannot control, and we provide a lightweight transferability test (Jaccard similarity against a partially-observed target-task graph) that separates successful from failed transfer in all four domain/structure pairs we evaluate, using under an hour of computation and 15-20% of target-domain data; we present this as a heuristic calibrated on few cases, not a validated general threshold.

0.46

Oracle-Efficient and Parameter-Free Agnostic Smoothed Online Learning

cs.LG stat.ML ·2026-10-07 ·Sasha Voitovych, Adam Block, Alexander Rakhlin, Abhishek Shetty
Abstract

Online learning is an attractive framework in many domains because it permits well-defined learning even when data are dependent or chosen adversarially. This generality, however, comes at a steep price, introducing significant statistical and computational barriers. Recently, smoothed online learning has emerged as a promising framework that interpolates between the fully adversarial and fully stochastic settings by assuming that the conditional law of each covariate has density at most $1/σ$ with respect to some fixed base measure $μ$, and it is known to match the statistical and computational guarantees of classical learning while still allowing for much of the flexibility of online learning. However, existing oracle-efficient algorithms require either (i) sampling access to the base measure $μ$ or (ii) labels that are perfectly predicted by a fixed hypothesis. Both assumptions limit the applicability of these algorithms, in contrast to statistical learning, where empirical risk minimization (ERM) learns efficiently in the agnostic setting without any knowledge of the data distribution. We show that neither assumption is necessary, giving the first oracle-efficient algorithm that achieves sublinear regret in the agnostic setting without knowledge of $μ$. Our algorithm, based on Gaussian Follow-The-Perturbed-Leader, is parameter-free: it requires no knowledge of $μ$, the smoothing parameter $σ$, or the horizon $T$, and it achieves regret $\widetilde O(d\sqrt{T/σ})$ for binary classes of VC dimension $d$ with a single call to an ERM oracle per round, which is optimal up to a $\sqrt{d}$ factor. En route to establishing the regret bound, we introduce several new techniques that may be of independent interest.

0.46

We Query, Therefore We Compute: On Oracle Computation beyond the Machine, with an Application to Agents

cs.AI ·2026-10-07 ·Kefan Liu, Fengning Ou, Yelin Luo, Jingdi Lei
code/data
Abstract · 86 pages, 11 figures, 17 tables

Agentic systems use large language models (LLMs) to carry out concrete tasks. Prior work often borrows abstractions such as scheduling, caching or isolation piecemeal from operating systems, so the mechanisms it builds share little common ground, and the shared view of the two forms of agentic system, Workflows and Agents, is limited. We construct an abstract machine that provides both. We treat the LLM as an Oracle and extend a two-stack pushdown automaton with one instruction, which hands the Oracle a whole stack as its query and appends the answer to that same stack. The machine thus performs two computations, the Oracle's and a Turing-complete one that we call the Priestess. A stack that the program only appends to grows autoregressively, as an agent's context does. Two symmetry breakings, S in storage and T in transitions, make a Priestess program the operating system of the programs the Oracle runs, and produce the Agent and the Workflow as the two placements of a task's program. For internally autoregressive Oracles, the two computations synchronize at the end of every answer under certain conditions, and through that synchronization we model caching and analyse scheduling. No guarantee that holds for every Oracle can fix which content crosses between the two computations, but such a guarantee does fix the boundary itself. The construction V fits the machine to a von Neumann computer. To show that it is realizable, we propose ArchNights, an extended RISC-V ISA and a Linux-style operating system implementing the machine by design. ArchNights-SE runs on gem5 as a computer system, becomes an agentic system when it runs an LLM as the Oracle, and will be open source. Agentic systems can then be designed as computer systems are. With a foundation built and a unified view, future work can share invariants and bounds, each with its conditions.

0.46

Explaining the Saliency Map Sparsity of Adversarially-Trained Neural Networks

cs.LG math.OCstat.ML ·2026-10-07 ·Yannick Lunk, Atell Yehor Krasnopolsky, Damien Garreau, Leon Bungert
theoryhighly cited
Abstract

Understanding why deep neural networks make a given prediction is of great importance for their safe deployment. In computer vision, saliency maps, which highlight the image region most influential for a prediction, remain a widely-used form of explanation. An empirical observation is the apparent sparsity of gradient saliency maps of adversarially-trained neural networks. In this paper, we propose a theoretical explanation of this phenomenon for two-layer ReLU networks. We build on the established equivalence of adversarial training to the minimization of the empirical risk with weight-decay penalization and an added adversarial total variation term -- valid for certain loss functions. As the number of data points and neurons grows and the regularization parameters are sent to zero at appropriate rates, we prove that minimizers converge to a Bayes classifier with minimal gradient and Barron norm. Sparsity appears since for adversarial training with $\ell_\infty$-attacks the gradient norm is anisotropic and favors axis-aligned / sparse gradients. We illustrate our theoretical findings experimentally by evaluating the gradient $\ell_1$-norm and thresholded sparsity of naturally versus adversarially trained models.

0.46

Budgeted Multi-Source Counterfactual Annotation for Off-Policy Evaluation

cs.LG cs.AImath.OC ·2026-10-07 ·Biao Xiang, Ali Eshragh, Yuexing Li, Kai Wang
top venuecode/data
Abstract · Accepted to NeurIPS 2026. Code available at https://github.com/Lygist/Budgeted-Multi-Source-Counterfactual-Annotation-for-Off-Policy-Evaluat…

Off-policy evaluation (OPE) estimates the value of a target policy from logged data, but limited behavior-policy coverage can force high-variance reweighting or reward-model extrapolation. Counterfactual annotations can add evidence about unobserved actions, yet practical sources, including domain experts and large language models (LLMs), may be costly, biased, or noisy. We study budgeted acquisition of such annotations for contextual-bandit OPE. Given source-specific costs and error profiles, we formulate an integer allocation problem over context-action pairs and annotation sources to minimize the component of estimator variance that depends on the annotation plan. We characterize when annotations are valuable through a first-annotation threshold and local annotation-value regimes. For the coupled multi-source problem, we develop a majorization-minimization algorithm with dynamic-programming subroutines that monotonically improves the objective. Experiments in synthetic clinical and LLM-annotated education bandits show that our allocation method reduces fixed-profile mean squared error (MSE) by 20.58% and 10.77%, respectively, relative to no annotation.

0.46

RSI-Forge: From Research Papers to Environments for Recursive Self-Improvement

cs.AI ·2026-10-07 ·Renxiong Wang, Darvin Yi, Abril Herrlein, Anas Mahmoud, Advait Gosai +14 more
Abstract

Environments are the foundation of recursive self-improvement: they provide the problems agents work on and the feedback used to evaluate progress. Yet constructing challenging research environments with reliable evaluation still depends on domain experts, limiting their scale and disciplinary coverage. We introduce RSI-Forge, a multi-agent pipeline that turns published papers into executable environments for self-improvement. Three agents coordinate construction, reproduction, and review to produce tasks with automated evaluators; each paper's method is independently reimplemented to establish a baseline score. We present 210 environments across 18 fields, including 90 reviewed by independent human domain experts. Both experts and agent judges give high ratings to the potential for improving the provided starting solutions and the evaluators' ability to distinguish solution quality, whereas experts are more critical of shortcut resistance, faithfulness to the source paper, and whether a single idea can exhaust a task. To validate their use for repeated improvement, we evaluate four models over 3 successive attempts on 120 environments, with each attempt inheriting prior code and notes while model weights remain fixed. At least one model improves after the first attempt in 84% of environments. Models also outperform the reproduced paper methods in 68 of the 120 environments, demonstrating room for gains beyond these baselines. Transcript analysis identifies work beyond parameter tuning in 95% of these successful attempts. Analysis of the resulting trajectories shows that models scoring lower on these tasks explore less, more often accept gains smaller than the reported standard error, and rely more heavily on tuning to the development set. RSI-Forge provides a scalable approach to constructing research environments for training and evaluating self-improving agents.

0.46

An extended deep energy method for thermo-mechanical crack propagation

cs.LG cond-mat.mtrl-scimath.NA ·2026-10-07 ·Han Zhang, Mehrisadat Makki Alamdari, Babak Shahbodagh, Mohammad Vahab, Cosmin Anitescu +2 more
Abstract · 48 pages, 20 figures, 10 tables

Thermo-mechanical fracture couples transient heat conduction on a cracked domain with a crack that grows as the temperature and the displacement evolve. Neural energy solvers have been proposed for phase-field fracture and later extended to represent a sharp crack through the network input, but heat conduction on the cracked domain and crack propagation under the resulting thermal stresses have not yet been treated together in these solvers. We present an extended deep energy method for thermo-mechanical crack propagation in which the crack remains a sharp polyline. Two networks represent the temperature and the displacement and receive the crack through a scalar embedding function, discontinuous across the crack and smooth elsewhere, so that both fields can jump across it without a regularization length, and the displacement is enriched near the tip by the Williams expansion with trainable amplitudes. The two fields are obtained by minimizing an incremental conduction functional and the thermoelastic potential energy in a staggered sequence, with Monte Carlo integration on points stratified over background elements, densified near the tip and redrawn during training. The stress intensity factors are extracted by the interaction integral with the area term of Wilson and Yu and checked by a sweep of the contour radius, and the crack advances at the maximum hoop stress angle when the energy release rate of the kink reaches the critical value at the crack-tip temperature. On a stationary thermal edge crack the extracted stress intensity factor agrees with the published value to 0.11%, in a functionally graded shear test initiation agrees with an independent sharp-crack finite element solution to within one load step, and on a notched cruciform specimen the crack paths follow the published solutions under mechanical, thermal and combined loading.

0.46

On the Clock: Towards Punctual and Productive Time-Budgeted AI Agents

cs.AI cs.LG ·2026-10-07 ·Aaron Wang, Neelabh Madan, Vlad Sobal, Matthew Trager, Michael Kleinman +3 more
published
Abstract · Accepted at the NeurIPS 2026 Workshop on Resource-Aware Agentic AI. 23 pages

We study whether small LLM agents can operate effectively under explicit wall-clock time budgets by both respecting the allocated runtime and using available time productively. We evaluate Qwen3.6-27B on five competitions from MLE-Bench Lite and Qwen3-4B on Zork I (Jericho), two agentic benchmarks where additional computational time can meaningfully improve performance. In the simplest setting, where the budget is stated only in the prompt, agents fail to translate the stated budget into controlled use of time. These failures arise from gaps in time awareness, since the harness provides no timing feedback, but also because they cannot reliably anticipate the duration of actions, and do not have a learned mapping from available time to an appropriate strategy. We investigate two complementary classes of interventions: harness-based mechanisms that expose timing information and enforce deadlines, and reinforcement learning with budget-aware rewards. Injecting timing information through the harness substantially improves budget adherence for Qwen3.6-27B without measurable loss in performance, while enforcement hooks tighten adherence further. RL with GRPO achieves near-perfect budget adherence on Zork I and generalizes to held-out budgets not seen during training, but does not improve task performance over the untrained harness on MLE-Bench. Once agents are made to respect the budget, they still fail to use additional time to improve task performance. RL-trained policies learn when to stop but often fill extra time with repeated actions, and GRPO training on multiple budgets tends to collapse toward the strategy learned for the shortest budget. Our results reveal a gap between time adherence and productive time allocation, which remains a central challenge for budget-conditioned agents.

0.45

Reliability of LLM Judges for Evaluating Entity Alignment

cs.AI ·2026-10-07 ·Vaibhava Lakshmi Ravideshik, Mayank Kejriwal
top venuebenchmarkcode/data
Abstract

Entity Alignment (EA) identifies equivalent entities across knowledge graphs and is critical for knowledge base integration and ontology merging. Evaluating EA systems at scale requires expensive expert annotation, making systematic assessment across diverse domains practically infeasible. LLM-as-judge evaluation offers a potentially scalable alternative, yet its reliability for structured prediction tasks like EA remains unstudied. We present the first systematic benchmarking study across three frontier models, three datasets, and four EA systems, using perturbation bias diagnostics, meta-evaluation across all dataset--judge--prompt combinations, and counterfactual label-flip tests. We identify anchor bias, a failure mode in which judges invert discrimination when the system's decision label is visible. Judge discrimination collapses under label-exposing prompts (J-ROC-AUC 0.12--0.87), while a label-free protocol recovers near-ceiling capability on distinctive-name datasets (0.98--1.00 for NeuSymEA) and significant recovery on biomedical pairs (0.93--0.95). Counterfactual label flips show that the visible label causally shifts scores by more than 2 points in 53--99% of pairs (FSR), with sensitivity growing with model capability, a pattern we call the frontier model paradox: stronger judges are more label-sensitive, not less, yet also better label-free predictors. The direction of the shifts is mixed, so we report this as an observed pattern. A blinded two-annotator human evaluation (102 pairs, Cohen's $κ$ = 0.902) provides a gold-label reference. We release the first biomedical EA benchmark (MeSH--SNOMED CT, 15K pairs) and a reproducible auditing framework for LLM judge reliability in EA. Code and data are available at https://github.com/vaibhavalakshmiravideshik/llm-as-a-judge-entity-alignment.

0.45

ResidualQuant: KV Cache Quantization for Looped Transformers with 2-Bit Residuals

cs.LG ·2026-10-07 ·Heejun Kim, Junyoung Lee, SangLyul Cho, Dongsu Han, Insu Han +1 more
highly cited
Abstract

Looped Transformers improve parameter efficiency by applying shared Transformer blocks over recurrent loops, increasing computational depth without increasing the parameter count. However, KV cache memory still scales with loop count, becoming a key memory bottleneck that limits batch size and inference throughput. KV cache quantization can alleviate this bottleneck, but existing methods often suffer substantial accuracy degradation at aggressive low-precision regimes. We observe that looped Transformers offer a unique opportunity: KV states across loops are highly similar. Based on this observation, we propose ResidualQuant, which uses the final-loop KV states as a reference and represents the remaining loops with low-precision residuals. Our method further combines least-square scaling and rotation-based quantization, as well as loop-wise mixed precision, to enable accurate quantization down to INT2 while retaining efficient reconstruction. Across multiple looped Transformer models and mathematical reasoning and code generation benchmarks, ResidualQuant consistently improves the accuracy-memory tradeoff over state-of-the-art rotation-based KV quantization. In particular, our method retains accuracy close to BF16 under mixed-precision settings while reducing theoretical KV storage by 80.7%, achieving up to 13.0% higher accuracy than the rotation-based baseline at nearly the same memory budget. On an RTX 5090, reduced KV memory traffic improves fixed-batch decode throughput by up to 3.66x, while the smaller memory footprint enables up to 8x larger batches, improving peak throughput by up to 5.38x.

0.45

CAFE+FNO: Fourier Kernel Generation via Multiplicative Feature Composition

cs.LG math.NA ·2026-10-07 ·Hyungjoon Juen, Minwoo Shin
code/data
Abstract

The Fourier Neural Operator (FNO) learns solution operators of partial differential equations (PDEs) through Fourier-space kernel parameterization, but frequency truncation can limit the learning of high-frequency variations. AM-FNO and SirenFNO generate kernels for all grid modes from spectral coordinates using shared networks, making coordinate encoding and generator design important. Recent work on implicit neural representations (INRs) has proposed constructing frequency interactions through explicit feature composition rather than relying on subsequent MLPs to form them implicitly. Building on this approach, we propose CAFE+FNO, which incorporates Content-Aware Frequency Encoding+ (CAFE+) into Fourier kernel generation. CAFE+ combines Fourier--Chebyshev features through parallel affine branches and a Hadamard product, forming interactions within and across the two feature families. A kernel MLP maps the resulting representation of each normalized spectral coordinate to a complex channel-mixing matrix. Each layer shares its generator across all stored modes, making the number of trainable parameters independent of the number of modes for a fixed architecture. We compare CAFE+FNO with existing FNO variants on five PDE benchmarks and conduct ablation studies on basis configuration, multiplicative composition, and bandwidth learnability. Code and experimental configurations are available at https://github.com/fabsk101/CAFEPlusFNO.git.

0.45

Evolutionary Architecture Search for Chlorophyll-$a$ Prediction in Lakes using Sentinel-2

cs.NE cs.LG ·2026-10-07 ·Kursat Komurcu, Linas Petkevicius
code/data
Abstract · Accepted at AutoML4EO 2026 (non-archival AutoML conference workshop). 4 pages + references. https://automl4eo.org/accepted-papers/

Small tabular datasets with expert-designed spectral features are the norm in operational Earth observation, and the networks applied to them are typically hand-designed. We revisit one such published model -- a Sentinel-2 algal bloom classifier -- and ask what architecture search adds, holding the task, the features and the lake-level train/test split of the original study fixed. Searching an extended multilayer-perceptron space with regularized evolution, and selecting on inner-cross-validation AUC only, we find networks that improve held-out AUC from 0.790 to 0.820 and accuracy from 0.733 to 0.748 while using 409 trainable parameters, 26 times fewer than the strongest hand-designed reference. The search converges on a consistent recipe -- a single narrow layer, RMS normalisation, $\tanh$ activation, step-decayed RMSprop and weight averaging -- that a practitioner would be unlikely to reach by default. At 1.6\,kB the resulting model is small enough to serve as an onboard screening trigger, which is the setting that motivates the work. Code: https://github.com/VU-AIML/automl4eo-bloom-nas.

0.45

How Do Agentic LLMs Decide to Call Tools? A Tool-Call Vector Shaped by Suppression

cs.AI cs.LG ·2026-10-07 ·Xijie Gong, Tingxu Han, Jiahao Zhang, Wei Song, Ziqi Ding +3 more
publishedcode/data
Abstract · Accepted by NeurIPS 2026 Main Poster

Tool calling, invoking external tools on demand, is central to agentic LLMs, yet the mechanism that decides whether a model calls a tool or responds directly remains poorly understood. Agentic prompts are long and heavily scaffolded, combining role instructions, tool schemas, format templates, and the user's request across hundreds of tokens, creating a noisy, highly entangled context in which no single controllable variable for mechanistic analysis is obvious. To obtain such a variable, we propose a method that converts complex agentic prompts into minimal contrastive pairs in which a single request verb determines the tool-call decision: replacing an execution-verb (e.g., \textit{write}) with an analysis-verb (e.g., \textit{discuss}) reliably flips the decision, suggesting it is mediated by a compact internal state. We construct 500 such paired prompts across Python, Java, and C++ (300 for mechanistic analysis, 200 held out for evaluation). We trace the decision to a vector, $μ_Δ$, that is both causally necessary and sufficient and generalizes beyond the discovery prompts to native multi-turn $τ^2$-Bench trajectories and verb-free requests. Behavioral ablations show that the scaffold establishes a tool-call prior; Transcoder decomposition then reveals that analysis verbs suppress this prior through features signaling that tool use is unnecessary, whereas execution verbs largely leave it intact. Downstream scaffold-reading attention heads and MLP features read out the resulting state, and the same mechanism recurs across seven models from the Qwen, Mistral, and Granite families. Our code is available at https://github.com/XijieGo/MI4ToolCalling.

0.45

Best Arm Identification for Bandits with Shifting Means

cs.LG stat.ML ·2026-10-07 ·Lukas Zierahn, Wouter M. Koolen, Shubhada Agrawal, Christina Katsimerou, Dirk van der Hoeven
top venuetheoryhot direction
Abstract · Accepted at NeurIPS 2026

We study the best arm identification problem in a stochastic environment with a novel form of adversarial perturbations, which we coin Shifting Means. While classically the mean rewards of the $K$ arms are stable in time, in Shifting Means only the gaps $\boldsymbolΔ$ between mean rewards are stable, while their common shift may be determined adversarially in each round. The objective of the learner is to identify the best arm with high probability while minimizing sample complexity (the fixed confidence setting). Handling shifts requires new tools: we show that algorithms employing a Generalized Likelihood Ratio Test (GLRT) stopping rule, including the popular Track-and-Stop, fail under time-varying shifts. Instead, we propose Importance Weights for Shifting Means ($\mathsf{ISM}$). Assuming means bounded by $U$ and $σ^2$-sub-Gaussian rewards, we show $\mathsf{ISM}$ to be $δ$-correct and to enjoy a sample complexity bound of order $K (σ^2 + U^2) Δ_{\min}^{-2} \ln \frac{1}δ$. We also present a matching (up to constant factors) worst-case lower bound and evaluate our results empirically.

0.45

EmbodiedRSI: Active Continual Robot Learning Through Hypothesis-Guided Co-Evolution

cs.AI ·2026-10-07 ·Python Song, Zhixuan Liang, Kelsey Fu, Mengdi Wang, Junfeng Yang +1 more
benchmarkhighly cited
Abstract

Robot foundation models provide strong visuomotor control, yet their performance can degrade when object positions or task instructions change. Further improvements often require post-training on substantial robot data, which can be costly to collect through methods such as teleoperation. Agentic harnesses can adapt around the model, but current self-evolving harnesses use robot trials inefficiently when deciding which code and skill changes to pursue. We introduce EmbodiedRSI, a self-evolving agentic harness that autonomously decides where to explore next and turns the resulting physical interaction into improved code and skills. EmbodiedRSI realizes this through a Fast-Slow Dual-System Architecture, in which competing code and skill hypotheses are maintained in a Hypothesis Graph. Value-of-Information Experiment Selection chooses physical experiments that can distinguish these hypotheses. Their outcomes guide Code-Skill Co-Evolution. The Slow System builds Hierarchical Memory, and Reward-Grounded Memory Learning selects effective memory according to their value for later Fast-System improvement. On RoboCasa365, EmbodiedRSI reaches 77.0% overall success and 71.3% on Composite-Unseen, compared with 40.1% for the best baseline. EmbodiedRSI also reaches 86.8% overall success on LIBERO-Pro. Beyond benchmark performance, EmbodiedRSI transfers zero-shot to real-world robot, achieving 71.3% overall success across multiple challenging tasks.

0.44

PatchBench: Measuring Collateral Damage in Activation Patching

cs.LG cs.CLcs.CR ·2026-10-07 ·Alexi Canesse, Mathis Le Bail, Maël Jenny, Clément Elliker, Mahammed El Sharkawy +1 more
top venuebenchmarkcode/data
Abstract · Accepted to NeurIPS 2026 (Datasets and Benchmarks Track)

An LLM safety patch can pass a benchmark while still being a poor repair. This risk is especially acute for jailbreak repairs, where the goal is to correct a specific unsafe behaviour without changing unrelated behaviours. A patch may block exact evaluation prompts yet fail on close harmful variants, or suppress harmful behaviour by over-refusing benign prompts that share its wording or structure. Existing protocols primarily test whether models can be broken, while aggregate metrics (attack success, refusal rates, global capability) cannot distinguish selective repairs from broader local suppression. To address this gap, we introduce PatchBench, a benchmark of empirically observed model-specific jailbreak failures inducing actionable harmful answers. Starting from 27,870 prompts from 37 public datasets, we curate 15,314 English prompts and query 8 open-source instruction-tuned models. Combining WildGuard filtering, pairwise Elo ranking, and manual verification, we retain a curated bank of 400 high-confidence jailbreak failures. We further introduce PatchBench-Local, an evaluation protocol testing whether a patch is behaviourally precise. For each harmful source prompt, PatchBench-Local generates three families of local neighbours: harmful variants preserving malicious intent, benign prompts with matched structure, and benign prompts reusing key harmful terms. It evaluates harmful-neighbour correction and benign-neighbour preservation, distinguishing selective repair from broader local suppression. Evaluating four activation steering methods with PatchBench-Local and MMLU shows that global capability can remain nearly unchanged while local benign regressions are severe, confirming aggregate metrics miss important collateral damage. PatchBench-Local provides a more precise basis for developing and comparing jailbreak repair methods.

0.44

Leaner Transformers Can Easily Learn to Cluster

cs.LG stat.ML ·2026-10-07 ·Charlotte Park, Kenneth L. Clarkson, Lior Horesh, Takuya Ito, Parikshit Ram
top venue
Abstract · NeurIPS 2026 accepted paper

Transformers have in-context learning capabilities, where some known learning algorithms can be executed in the forward pass through the model. Recent work shows that transformers can exactly perform Lloyd's algorithm for $k$-means clustering with $n$ points in $d$ dimensions with an embedding size $d_{\textsf{emb}} = d+k$ (thus, requiring attention projection matrices of size $(d+k)^2$). In this work, we build upon this result in the following ways: First, we present an equally expressive but smaller transformer that executes Lloyd's algorithm with embedding size $d_{\textsf{emb}} = (d + \lceil \log_2 k \rceil)$. Next, we train these transformers to learn the clustering algorithms given a distribution of clustering tasks, and theoretically characterize and empirically validate the factors affecting the convergence and in-distribution generalization of learning algorithms based on stochastic gradients. Finally, we probe the general clustering abilities of these learned algorithms (in the form of transformers), and try to understand situations where they succeed and fail.

0.44

The Modular CMA-ES: A Framework for Modern Evolution Strategies

cs.NE ·2026-10-07 ·Jacob de Nobel, Diederick Vermetten, Anna V. Kononova, Carola Doerr, Thomas Bäck
benchmark
Abstract

Since their introduction, modern evolution strategies, such as the Covariance Matrix Adaptation Evolution Strategy (CMA-ES), have become established as powerful methods for continuous black-box optimization. This success has led to a wide range of proposed modifications, each designed to improve performance or behavior in specific optimization scenarios. However, because these developments have largely been introduced and studied in isolation, their interactions remain comparatively underexplored. In this paper, we present the Modular CMA-ES (ModCMA), a configurable framework that integrates a wide range of mechanisms from modern evolution strategies within a single implementation. By decomposing CMA-ES into modules with interchangeable options for sampling, selection and recombination, step-size adaptation, matrix adaptation, and restarting, ModCMA enables systematic exploration of a large design space of modern evolution strategies and facilitates the construction, comparison, and automated configuration of new algorithm variants. We illustrate the benefits of this modular approach through two example studies. First, we compare several matrix-adaptation mechanisms in terms of their computational cost and optimization performance. Second, we use automated algorithm configuration to specialize ModCMA to individual benchmark problems and analyze the resulting configurations. Together, these examples demonstrate how the framework can be used both to study individual algorithmic design choices and to explore their combinations in a systematic and reproducible manner.

0.43

Bolzano: From Expert-Guided Proof Search to Automated Open-Problem Solving

cs.AI cs.CL ·2026-10-07 ·Adrián Zámečník, Matěj Kripner, Martin Koutecký, Martin Balko, Jan Grebík +3 more
publishedcode/data
Abstract · Accepted at the 6th Workshop on Mathematical Reasoning and AI (MATH-AI), NeurIPS 2026

Large language models are increasingly contributing to mathematical research, where progress often depends on efficient proof search, incremental improvements and careful verification. We describe Bolzano, a multi-agent open-source system that uses parallel prover agents with a verifier agent and maintains a human-readable research state. Initial manual use on expert-selected problems yielded 8 results whose proofs were checked by domain experts. Motivated by these case studies, we ran Bolzano without problem-specific human guidance on about 3,800 open problems extracted from four sets of papers, solving about 200 open problems. One experiment used papers accepted to STOC 2026, a top conference in theoretical computer science. There, we answered four questions raised in the papers, as confirmed by their authors.

0.43

Loud Failures, Quiet Failures: Fault Detection and Recovery in Tool-Using Language Model Agents

cs.AI ·2026-10-07 ·Obada Kraishan
benchmark
Abstract

Tool-using agents are usually scored on whether they finish a task while the tools work. Deployments are less forgiving: services time out, endpoints disappear, parameter names change, and results come back well formed but wrong. Prior work has shown that language models over-trust tool outputs that fail silently; we ask how that over-trust plays out across the stages of failure handling in multi-turn agents. Wrapping the executable environments of an established function-calling benchmark in a fault-injection layer, we inject one of four typed faults at a controlled point in the trajectory and record whether the agent notices, changes plan, recovers the task, or repeats itself. Six models from three families, half of them reasoning variants, ran 1,920 trials over 24 multi-step tasks. Agents treat a failure as a problem in 91.3% of trials when the tool returns an explicit error, but in 58.8% of trials when it returns a plausible wrong value, against a 26.8% rate of reporting problems when nothing was wrong. Reasoning models are not better placed: paired against instruct siblings, they notice less (-9.3 points, p < .001) and change plan more (+10.4 points, p < .001), and recovery is unchanged (p = .512). Because agents are stochastic, two fault-free runs of the same task end in the same state only 63.3% of the time; against that baseline, only a missing tool clearly lowers recovery (39.9%), while timeouts, schema drift, and corruption stay within run-to-run variation. After a fault, agents return to the same tool three or more times in a row in up to 22.2% of trials, though strictly identical repeats are rare. A prompt line asking the agent to check each result did not move detection. Agents respond to the error channel rather than to the content of what a tool returns, so failures that stay inside the expected format pass through.

0.43

Learning to Accumulate Knowledge with Mutual Information

cs.AI ·2026-10-07 ·Yuyang Zhao, Lizi Liao, Leyang Shen, Xiaoyan Zhao, Yang Zhang +2 more
code/data
Abstract

Large language model (LLM) agents can improve their performance by reusing knowledge distilled from past interactions. However, curating new experiences into a knowledge bank that becomes more useful as it grows remains challenging. Effective knowledge accumulation should limit redundant overlap among entries and ensure that new knowledge contributes beyond what the bank already provides. Yet training a curator with Group Relative Policy Optimization (GRPO) on standalone task success can reinforce general guidance even when it duplicates existing knowledge. Therefore, we propose Knowledge Weaver, a reinforcement learning framework that trains a language model to curate reusable knowledge from agent trajectories. We couple feedback inspired by token-wise mutual information (MI) with marginal success rewards to guide knowledge accumulation. Together, these signals encourage the curator to preserve distinct information from experience and produce entries that improve task success when added to existing knowledge. Standalone success rewards also favor entries that are useful on their own. On ALFWorld and WebShop, Knowledge Weaver achieves mean success rates of 54.0\% and 42.0\% with k=10 retrieved entries, exceeding GRPO by 16.9 and 18.7 percentage points, respectively. Its knowledge banks also outperform the evaluated prompt-based and established banks, including human-written banks, in overall ALFWorld success rate and WebShop score with the executor frozen. Our codebase is available at https://github.com/LaoKuiZe/Knowledge-Weaver.

0.43

Progress and Prospect of AI in ARPES Workflow

cs.LG cond-mat.otherphysics.data-an ·2026-10-07 ·Sandy Adhitia Ekahana, Aalok Tiwari, Pratik Saud, Aaron Bostwick, Chris Jozwiak +2 more
code/data
Abstract

Artificial intelligence (AI) is becoming an increasingly useful tool across the experimental sciences, including angle-resolved photoemission spectroscopy (ARPES), which routinely produces large, multidimensional datasets of electronic structure. Recent advances in AI and machine learning (ML) have opened new opportunities across the entire ARPES workflow, from automated sample preparation and real-time data acquisition to post-experiment data analysis and comparison with theoretical calculations. Despite this progress, a comprehensive review of ML applications, their capabilities, and reliability across the different stages of ARPES workflow is still lacking. In this review, we first introduce ML methods that are most relevant to experimentalists working in condensed matter physics and materials science. We then follow the ARPES workflow, reviewing existing ML applications at each step and discussing their advantages, limitations and potential for future development. We also examine the current ARPES data landscape, where several open databases are available but remain relatively small and fragmented compared with large, shared datasets such as ImageNet. Given these limitations, we suggest that the community focus on sharing pretrained models that can be further trained, adapted to specific tasks, and redistributed, while working toward a larger and standardized open ARPES dataset repository. Finally, we discuss our perspectives on the future of AI within the ARPES workflow using a six-level framework of laboratory automation, highlighting the opportunities and challenges in moving toward a fully autonomous, self-driving ARPES laboratory.

0.43

Temporally Interpretable Differentiable Decision Trees

cs.LG cs.RO ·2026-10-07 ·Eisuke Hirota, Aarav Sane, Rohan Paleja
code/data
Abstract

Interpretability offers a solution to safe autonomy by providing transparency into an agent's underlying decision-making model. Within sequential-decision making tasks, differentiable decision trees (DDTs) are one approach to such interpretability, maintaining automatic-differentiable policies while providing humans with a discrete tree-based visualization. Nonetheless, current implementations of DDTs are not well-suited for sequential-decision making domains, as there exists an inherent mismatch between a tree's single-timestep behavior and a human's multi-timestep planning. Our work thus introduces time as a new dimension of interpretability, coined as temporal interpretability, and demonstrates how temporal abstractions via action chunking improve it. We achieve this by first introducing two novel policy gradient algorithms that incorporate action chunking. Additionally, to maintain parameter-efficient trees, we develop an information-theoretic tree restructuring algorithm that modifies the tree during training. Across four simulation environments, we find that warm-starting action chunked DDTs from a distilled action chunked policy is the most effective way to obtain temporally interpretable trees: they match neural network policies in three of the four domains while using up to 80$\%$ fewer parameters. Our code is available at https://github.com/ei5uke/temp-interp.

0.43

Measurement-Efficient Differentiable Quantum Architecture Search for Combinatorial Optimization

cs.LG quant-ph ·2026-10-07 ·Lukas Theißinger, Thore Gerlach, Christian Bauckhage
top venuebenchmarkcode/data
Abstract · 6 pages, 2 figures. Accepted at the 2026 IEEE 2nd International Conference on Quantum Artificial Intelligence (QAI). Code and data: https://…

Differentiable quantum architecture search (DQAS) is a promising framework for the automated design of quantum circuits, particularly for variational quantum optimization algorithms. However, its practical deployment on quantum hardware is limited by the large number of circuit measurements required during optimization, making hardware execution costly. In this work, we show that for a broad class of combinatorial optimization problems and commonly used rotational gate parameterizations, the measurement cost of DQAS can be significantly reduced without changing the optimization objective. We derive the proposed measurement reduction scheme theoretically and validate it experimentally on 3-SAT and MaxCut benchmark problems. Our approach reduces the requested gradient measurement cost by about 39 to 41% while introducing only negligible classical post-processing overhead, lowering the practical cost of executing DQAS on quantum hardware.

0.43

MovieSTAGE: Scene, Transition, and Global Encoding for Movie-fMRI ADHD Classification

cs.LG q-bio.NC ·2026-10-07 ·Boseong Kim, Haejun Chung, Ikbeom Jang
published
Abstract · Accepted to the 2026 IEEE International Conference on Bioinformatics and Biomedicine (BIBM 2026)

Naturalistic movie-fMRI provides a shared, temporally structured probe of brain dynamics, yet predictive models commonly rely on whole-run functional connectivity (FC) or temporally generic representations that are not aligned with narrative events. We introduce MovieSTAGE (Scene, Transition, and Global Encoding), a multiscale framework that combines hypergraph-structured FC-profile organization within scenes, unsigned FC-profile differences across adjacent scenes, and whole-movie FC. We evaluated 260 participants from the CMI-HBN Despicable Me cohort on case-control, ADHD-subtype, and three-class classification using 10 repetitions of stratified five-fold cross-validation, complete out-of-fold (OOF) predictions, and paired subject-cluster bootstrap and permutation tests. MovieSTAGE achieved AUROCs of 0.69, 0.73, and 0.75 and balanced accuracies of 67.6%, 69.8%, and 58.3%, respectively, yielding the highest mean point estimates among the evaluated methods. On the three-class task, the full model outperformed all two-branch variants, the HGNN scene encoder outperformed MLP, GAT, and BNT alternatives under matched settings, and the human-annotated partition outperformed duration-matched random and fixed-count GSBS controls. These controlled results support incremental predictive value from event-aligned scene and transition representations when combined with whole-movie FC in this cohort. Post-hoc model-derived analyses generated network-level hypotheses involving frontoparietal and default-mode systems.

0.43

Cross-Domain Pretraining for Steady-State Neural CFD Surrogates

cs.LG ·2026-10-07 ·Anthony Zhou, Amir Barati Farimani, Shirley Ho, Rudy Morel
dataset
Abstract · 43 pages, 25 figures

Neural surrogates for computational fluid dynamics (CFD) have the potential to greatly enhance engineering innovation through accelerating simulation. However, the primary limitation for neural surrogates is the lack of generalization to geometries and applications beyond the training set, which is significant given the diversity of engineering scenarios. Currently, this is addressed by generating a new dataset for a specific application; however, this requires running costly numerical solvers. In this work, we take a step toward addressing this by studying neural surrogates trained across different geometries, boundary conditions, and fidelities. We find that cross-domain pretraining improves zero- and few-shot performance on held-out datasets relative to both training from scratch and transferring from domain-specific experts. In particular, finetuning a pretrained, cross-domain model can achieve 2-3x lower errors at the same sample size and use 8x fewer samples to achieve the same error, compared to training from scratch. This benefit is architecture agnostic and improves with model size and pretraining dataset diversity. Furthermore, we study how and why cross-domain pretraining works in CFD surrogates, and find that simply pooling steady-state datasets is both sufficient and effective. Given the high cost of generating CFD data, leveraging existing datasets through cross-domain pretraining will likely be a valuable strategy as future surrogates expand to tackle new problems and use cases.

0.42

HGP:An on-device personalized agent memory via hybrid graph storage

cs.AI ·2026-10-07 ·Ran Zhou, Xueming Han, Jiaheng Liu, Yuyao Zhang, Fanyu Meng +2 more
top venuecode/data
Abstract

LLM-based agents face challenges in personalized interactive tasks due to heterogeneous, multi-typed, and implicitly constrained long-term traces. Existing memory mechanisms struggle with accurate routing and retrieval, especially on-device where personalization is critical. Most methods use single-vector representations, blurring type distinctions and relational structure. We propose HGP, a hybrid graph memory framework. HGP employs a lightweight self-enhancement classifier for personalized memory routing and constructs episodic, semantic, and procedural memories as graphs. It also extracts working memory as a state trajectory to capture current state and implicit constraints, ensuring reliable decision-making. The classifier reduces large-model calls, enabling on-device deployment, while graph storage enables accurate retrieval and incremental user profile refinement. Experiments on two benchmarks show that on PAL-Set solution selection, HGP achieves an S-score of 35.58, nearly 7 points above the strongest baseline. Code and data are at https://github.com/Ouan6/HGP-.git.

0.42

RH-Detect: A Unified Benchmark for Reward Hacking Detection

cs.LG ·2026-10-07 ·Junwei Quan, Evgenii Opryshko, Rohan Subramani, Igor Gilitschenski
benchmark
Abstract · 16 pages, 3 figures, 16 tables

Reward hacking, where a model exploits an evaluation signal without completing the intended task, threatens the reliability of deployed language model systems. Existing datasets use different labels, response formats, and metadata conventions, making detector results difficult to compare. We present RH-Detect, a benchmark that combines reward-hacking-relevant subsets from eleven public datasets, comprising 92,761 rows and six behavior categories, into a common schema. On 5,021 open-ended evaluation units, each comprising a task prompt and a free-form model continuation, including multi-turn tool-use trajectories, we evaluate six off-the-shelf language models from five families as reward hacking detectors without additional training. The best model achieves a pooled AUROC of 0.962, with accuracy above 93%. For the four strongest models, however, accuracy on the two multi-turn tool-use datasets, MALT and TRACE, is 10.7-15.9 percentage points lower than on the other sources at a common decision threshold, highlighting a key gap for deployment-time monitoring. We find that different input formats have different effects across models. Removing thinking raises Qwen3.5-4B AUROC from 0.779 to 0.849, but lowers Qwen Flash from 0.977 to 0.950. We also evaluate the benchmark as a training dataset for detectors. Holding out each source in turn, single-token SFT improves average AUROC on five of six held-out sources. A GRPO follow-up on that failure case yields a slight improvement in detection performance. Our results show that a single pooled score can conceal variation across data sources, detector inputs, and training procedures.

0.42

Symbolic Density Estimators for Unnormalized Distributions

cs.LG hep-lat ·2026-10-07 ·Vikas Kanaujia, Riyansha Singh, Shashank Sharma, Vipul Arora
top venue
Abstract · 32 pages, 3 figures

Estimating the symbolic or analytical form of probability density functions (PDFs) from observed samples is a fundamental challenge in statistical and computational modelling. This process is critical for deriving interpretable and generalizable relationships characterizing the underlying phenomenon. Traditionally, this estimation depends strongly on domain expertise and prior field-specific knowledge, with experts selecting appropriate functional forms or parametric families based on empirical evidence and theoretical understanding. The coefficients of these forms are then typically determined through parameter estimation. In this paper, we develop a framework for estimating symbolic expressions of unnormalized distributions from observed samples using domain-specific prior knowledge, such as the range of interactions and a predefined set of primitive functions. We integrate deep generative models with symbolic regression (SR), incorporating inductive biases, such as factorizing large distributions, to keep the problem tractable. The deep generative models we examine include likelihood-based models, viz., flow models, and score-based models. Experiments show the effectiveness of the proposed framework for estimating density functions for multivariate toy distributions as well as lattices from computational physics, namely, XY model and $φ^4$ theory. When applied to the renormalization problem in $φ^4$ theory, the proposed framework estimates compact symbolic approximations of the hamiltonian function at different scales directly from samples, yielding expressions that may be challenging to derive using traditional perturbative or analytic approaches in nonperturbative settings.

0.42

F$^3$NO: Frequency-Decomposed Finite-Time Flow-map Neural Operators with Cross-Scale Conditioning

cs.LG ·2026-10-07 ·Fan Wu, Cheng Jing, Kookjin Lee
Abstract

Neural operators enable fast PDE forecasting, but repeated predictions accumulate errors and fine-scale structures remain difficult to resolve. We introduce a frequency-decomposed finite-time flow-map neural operator (F$^3$NO) that leverages updated low-frequency features to guide nonlinear refinement of high-frequency information. Within each layer, this cross-scale conditioning connects global spectral processing with local detail refinement. The model directly predicts states at specified future times and adjusts the contributions of the two branches according to the prediction interval. For longer trajectories, it combines parallel predictions within short temporal segments with recursive propagation between segments. Experiments on five PDE benchmarks demonstrate improved forecasting accuracy over autoregressive and direct-prediction baselines. Ablations show that frequency-decomposed refinement can improve accuracy with fewer parameters, while the benefits of segmentation depend on spatial resolution and dynamical regime.

0.42

Decentralized SGD under Heavy-Tailed Noise: Optimal Convergence Rates and the Role of Gradient Clipping

cs.LG math.OCcs.MA ·2026-10-07 ·Aleksandar Armacki, Haoyuan Cai, Ali H. Sayed
Abstract · 36 pages, 5 figures, 2 tables

Heavy-tailed noise has been widely observed in modern machine learning, motivating the use of methods like gradient clipping and normalization. While these methods are well understood in centralized settings, much less is known in decentralized ones, where applying a nonlinearity to local gradients affects both optimization and consensus. Recent works on decentralized non-convex optimization have studied both clipping and normalization under heavy-tailed noise, with clipping yielding suboptimal rates and normalization needing local momentum or mini-batches to converge. This raises the question: can a baseline decentralized method using a nonlinearity achieve optimal convergence rates under heavy-tailed noise? We answer affirmatively with clipped decentralized SGD ($\mathtt{DSGD}$). For smooth non-convex costs under bounded $p$-th moment noise, $p \in (1,2]$, we show that clipped $\mathtt{DSGD}$ achieves order-optimal rates both with high probability and in expectation. Moreover, we establish a linear speed-up in the number of agents, which, to our knowledge, has not been shown for decentralized methods with clipping. The key technical ingredient is a sharp analysis of the consensus gap that exploits the structure of clipping, relegating network effects to higher-order terms. Our results highlight an important distinction between clipping and normalization in decentralized settings: while normalized $\mathtt{DSGD}$ can fail to converge, clipping retains magnitude information, enabling $\mathtt{DSGD}$ to be convergent and order-optimal. Numerical experiments validate our theory.

0.42

Similar Predictive Fit but Different Latent Dynamics: Characterizing Learned Dynamical Structure in Personalized Models of Brain Disorders

cs.LG eess.SPq-bio.NC ·2026-10-07 ·Rita Huan-Ting Peng, Nhat Bui
published
Abstract · Accepted at the World Models for High-Stakes Health (WMHS) Workshop at NeurIPS 2026

As AI models move toward clinical decision-making and personalized treatment, understanding \emph{what} a model learns is important beyond predictive accuracy alone. We investigate whether personalized latent dynamics reveal clinically associated differences even when predictive fit is similar. A lightweight CNN--Transformer EEG foundation model pretrained on the Temple University EEG Corpus (TUEG) extracts segment-level representations. Using the Temple University Epilepsy Corpus (TUEP), representations are mapped to a shared latent-state space, and sparse multinomial logistic transition distributions (mLTD) are fit independently to each subject to obtain personalized transition-dependency graphs $W_n$. Analyses include $n{=}198$ subjects (99 epilepsy / 99 non-epilepsy). At $k{=}4$, epilepsy subjects exhibit substantially denser learned dependency structure ($p{=}1.1\times10^{-7}$), with the same pattern at $k{=}6$ (19.90 vs. 13.46; $p{=}5.2\times10^{-5}$). Graph-derived features provide moderate group discrimination under 5-fold subject-wise cross-validation (AUROC 0.68 at $k{=}4$; 0.65 at $k{=}6$). In contrast, held-out log-likelihood is nearly identical between groups at $k{=}4$ ($-0.992$ vs. $-0.991$; $p{=}0.95$), with similarly matched next-state prediction (AUROC 0.855 vs. 0.861; $p{=}0.54$). Thus, similar predictive fit does not imply similar learned dynamics: groups can be comparably predictable while differing substantially in the internal dynamical structure learned by personalized models. This distinction motivates evaluating learned structure alongside predictive performance in personalized clinical models.

0.41

DrugTargetWorld: A Synthetic Biobank for Training and Benchmarking AI Scientists

cs.AI ·2026-10-07 ·Samuel Margolis, Paul Schmiedmayer, Alan Huang, Ethan Chen, Ishan Bhattacharjee +10 more
code/data
Abstract · 34 pages main text, 92 pages supplementary material; 6 main figures. Project: https://DrugTargetWorld.vercel.app. Code: https://github.com/s…

Drug target discovery requires distinguishing molecules that causally drive disease from those that are merely associated with it. Training and evaluating AI agents to perform this workflow end-to-end is difficult because real world biobanks lack known causal ground truth and participant-level data is access controlled. We introduce DrugTargetWorld, a framework that procedurally generates simulated biobanks, or "worlds," with known but concealed causal structure. Each world contains genotypes, proteins, health records, outcomes, and synthetic magnetic resonance imaging (MRI) for 54,000 participants. Agents must construct a disease phenotype, identify causal driver proteins, infer the beneficial direction of modulation, and optionally conduct virtual 'wet lab' experiments. We evaluated nine agents in 540 episodes across 20 cardiovascular worlds and three experimental budgets. Opus 5 and GPT-5.6 Sol achieved the highest mean composite scores, 39.98 and 35.38 of 100, respectively, and both recovered 64% of causal drivers on average. However, no agent reliably distinguished misleading non-causal proteins, and performance remained limited by the integrative judgments required to connect phenotype construction, causal evidence, and intervention decisions. By making each world's causal structure known to the evaluator but hidden from the agent, DrugTargetWorld turns end-to-end drug target discovery into a scalable training and evaluation problem with verifiable reward.

0.41

RollVerify: Bridging Efficiency and Accuracy in Long-Tail Rollout Reinforcement Learning

cs.LG cs.AI ·2026-10-07 ·Yongqiang Yao, Jinru Tan, Kaihuan Liang, Zixin Yin, Yazhe Niu +3 more
Abstract

Reinforcement learning is crucial for improving large language models' reasoning and generalization. It relies on massive rollouts whose lengths become increasingly long-tailed as context windows grow. In on-policy training, these long-tail rollouts can result in GPU bubbles, reducing system utilization and limiting RL scalability. Asynchronous or partial-rollout methods improve throughput by relaxing synchronization, but inevitably introduce stale off-policy samples (trajectories) that may hurt final accuracy. Existing approaches mainly mitigate this off-policy issue by reweighting off-policy samples during training, yet they can still leave a performance gap compared to fully on-policy training. In this work, rather than passively reweighting samples during training, we propose RollVerify, a lightweight RL framework built on partial rollout that actively verifies and repairs samples before they enter training. Specifically, it introduces an off-policy shift metric OPS, to quantify the off-policy deviation of partially generated trajectories. Guided by the OPS constraint, RollVerify performs both sequence-level and token-level verification to identify and truncate invalid suffixes of trajectories. This yields high-quality samples that protect the models' accuracy while preserving the efficiency gains of partial rollout. Experiments on mathematical and tool-assisted mathematical reasoning show that RollVerify achieves accuracy comparable to on-policy training while reducing training cost. Additional code-generation results provide preliminary evidence beyond mathematics.

0.41

SciExam for ENSO: Can AI Agents Build Climate Models?

cs.AI cs.LGphysics.ao-ph ·2026-10-07 ·Yinling Zhang, Langchen Liu, Dongbin Xiu, Xueyan Zou, Xu Kuang +2 more
benchmarkcode/data
Abstract · 28 pages, 5 figures, 8 tables. Code: https://github.com/ylzhang2447/SciExam-ENSO-code

Language-model agents are increasingly asked to carry out open-ended scientific research, yet their results are usually graded against a known answer, a rubric, or a language-model reviewer, none of which can tell whether a new scientific model is valid. The AI Science Exam for El Nino-Southern Oscillation (SciExam for ENSO) is a benchmark in which agents build low-order stochastic models of ENSO, the dominant mode of interannual climate variability, from real observations. Within a six-hour budget, agents process the observations, write their own diagnostics, which are then frozen, and develop a model using only these diagnostics as feedback. Hidden graders then test whether the model reproduces ENSO's statistics, recovers unobserved variables, and forecasts held-out years, and score a published model in the same way. Across twelve agent systems, six produce models that score higher than the published model, mainly through better reconstruction and forecasting. The simplified forms of the stronger models are each compatible with one of the two competing explanations of ENSO's warm-cold asymmetry, an open debate that the task never mentions. Controlled runs of the top system under varied information suggest that its scores do not come from recalling the dated observational record and that the information it receives shapes how it builds its model. SciExam for ENSO can thus evaluate agent research where no answer is known, and the results suggest that agents can already build competitive models whose structures bear on questions that scientists still debate.

0.41

Scalable Logistic Gaussian Process Density Regression with Kinetic Langevin Sampling

cs.LG stat.ML ·2026-10-07 ·Daniel Paulin, Ádám Jung, András A. Benczúr
hot direction
Abstract · 27 pages, 3 figures

Conditional density estimation targets the full distribution of a response given covariates, as required, for example, for per-galaxy photometric redshifts. We develop a scalable Bayesian estimator based on the logistic Gaussian process. The log conditional density has a separable covariance: a Matérn kernel along the response, represented in a truncated Fourier basis on a circle, and a covariate kernel represented by Nyström features, which accommodate non-stationary kernels with input-dependent amplitudes and length scales. Instead of a Laplace or variational approximation, we sample the latent field of this finite-feature model. Given the hyperparameters, its posterior is strongly log-concave with a uniformly bounded Hessian, and we draw from it by simulating kinetic Langevin dynamics with symmetric minibatch splitting in Kronecker-whitened coordinates. Marginal-likelihood gradients follow from Fisher's identity as posterior expectations. Under the conditions of our analysis their bias is controlled by the sampler's step size and run length, and the predictive averages over the non-Gaussian latent posterior instead of a Gaussian around its mode. On photometric-redshift benchmarks with up to 3.9 million training observations, trained on a single GPU, the estimator is competitive with state-of-the-art tabular foundation models on density and calibration metrics.

0.41

From Prompts to Trees: Effective LLM-Guided Tree Generation for Few-Shot Tabular Classification

cs.LG cs.AIcs.CL ·2026-10-07 ·Yue Qiu, Zekang Du, Yiqun Diao, Bingsheng He, Qinbin Li
top venuecode/data
Abstract · Accepted to EMNLP 2026 Main as an oral presentation. Code available: https://github.com/yueqiu0/LLMTree

While Large Language Models (LLMs) possess rich world knowledge and impressive generalization capabilities, their direct application to tabular data classification is hindered by high inference costs and limited interpretability. In contrast, decision trees are fast and transparent but often underperform in low-data regimes. In this work, we propose a novel framework that bridges these paradigms by distilling LLM knowledge into interpretable decision trees under a few-shot learning setting. Instead of directly prompting the LLM to generate full trees, which is often unstable and inefficient, we develop a three-stage paradigm that prompts the LLM to generate rules and organize the rules into a tree. Experiments on multiple real-world tabular datasets demonstrate that our method achieves superior accuracy and interpretability with significantly lower prompting overhead compared to existing baselines.

0.40

CANDO: Cooperative Agentic Network for Layout Design Optimization

cs.MA ·2026-10-07 ·Athanasios Masouris, Zheng Jing, Benjamin Sam Chandler, Hadi Jamali-Rad
top venuebenchmark
Abstract · Accepted to appear in NeurIPS 2026. This is the preprint version of the paper

Layout generation for real-world facilities is a challenging problem, requiring reasoning over irregular site boundaries, heterogeneous orientations, access-aware placements, and motion-planning feasibility. Yet, most existing layout benchmarks in the generative AI space target simpler placements over rectangular domains and rely on distributional metrics such as FID and IoU that reward conformity to dataset priors, thus discounting design innovation. Motivated by these gaps, we introduce ALPS-Bench, a benchmark of $1,000$ professionally annotated real-world facility layouts paired with an instance-specific scoring protocol grounded in a structured design manual. As a strong baseline for ALPS-Bench, we propose CANDO, a training-free multi-agent framework in which specialized agents iteratively refine layouts through a verification-grounded loop, concentrating reasoning on strategic spatial decisions. We demonstrate that CANDO surpasses state-of-the-art trained and LLM-based baselines on the widely adopted PubLayNet, RICO, and PKU-PosterLayout benchmarks, establishing cooperative agentic design as a broadly effective recipe for constraint-aware layout synthesis.

0.40

COPC: Coupled Off-Policy Correction for Asynchronous LLM Reinforcement Learning

cs.LG ·2026-10-07 ·Zicheng Hu, Zhijian Zhou, Xuan Zhang, Yuchen Liu, Cheng Chen +5 more
Abstract

Asynchronous RL accelerates large language model post-training by decoupling rollout generation from optimization, but trains on stale trajectories. Existing methods primarily correct token-level policy mismatch through importance-ratio control in the actor objective. We show that this \emph{policy-side correction} alone is insufficient: advantage estimates also inherit mismatch from behavior-policy continuations, which we term \emph{advantage staleness}. We derive exact bias and variance decompositions for a general two-channel actor update, revealing nonseparable coupling between policy-weight and advantage-estimation errors: their interaction induces multiplicative bias terms, while squared policy weights amplify advantage uncertainty in gradient variance. This motivates the hypothesis that policy- and advantage-side correction should be coordinated. We introduce Coupled Off-Policy Correction (COPC), an actor--critic method combining token-level ratio masking with two-sided clipped-ratio weighting of TD residuals for return and advantage estimation. Joint parameter sweeps across staleness levels support this hypothesis: the effect of one correction parameter depends on, and can reverse with, the other. COPC achieves the highest reported performance on tool-integrated mathematical reasoning and search, outperforming the strongest reported asynchronous baseline in each setting. It also offers a broad high-performing parameter region and improved training stability. In search, COPC remains stable throughout training, while most evaluated asynchronous baselines collapse late in training. These gains persist at 64-step policy staleness. COPC adds minimal step-time overhead over asynchronous PPO and retains a $1.7\times$ step-time speedup over synchronous PPO.

0.40

AgentTime: Can Agents Estimate and Control Their Own Runtime?

cs.AI ·2026-10-07 ·Michael Ofengenden, Maksym Andriushchenko
benchmarkcode/data
Abstract · Website: https://agenttimebench.com Code: https://github.com/michaelofengenden/agenttimebench

An essential control of AI agents is their ability to manage runtime. This ability requires a sense of time-awareness, to predict and estimate wall-clock time and to control their own actions. Prior work has focused on time-awareness, but duration-following and control in native agent harnesses remain unexplored. We present AgentTime, a benchmark for testing whether agents can work for a requested duration, predict their runtime, and estimate elapsed time afterward. It comprises 222 tasks from 18 sources spanning coding, computer use, agentic work, and automated research. Duration-following experiments append a single instruction specifying how long to work, with requests ranging from about a minute to multiple days. Accuracy on these instructions varies substantially: Fable 5.1 in Claude Code deviates from requested runtimes by a typical factor of 2.9$\times$, compared with only 1.2$\times$ for GPT-6 Astra in Codex. However, matching the requested runtime does not, by itself, establish continued work on the task. Among 158 reviewed Astra runs with classifiable transcripts, 14 explicitly slept after appearing to finish. In forecasting experiments, predictions tend to overestimate natural runtimes. In retrospective experiments, removing temporal information more than doubles deviation for Sol and Astra and nearly doubles it for Fable. An agent's ability to complete a task does not guarantee that it can control its own time or work for the whole requested duration. For agents to run reliably, safely, and autonomously over long horizons, we require the evaluation of both.

0.40

RoboJEPA: Scaling Robotic Latent World Models

cs.AI cs.RO ·2026-10-07 ·Artem Zholus, Nicolas Beltran-Velez, Jianhao Yuan, Sarath Chandar, Tushar Nagarajan +7 more
Abstract

Latent world models have shown a remarkable ability to predict future states and to plan in the real world. In practice, however, we lack a principled way to estimate how their capabilities scale with model size, data, and compute, an open problem that slows progress in the field. In this work we present RoboJEPA, a world model based on the Joint Embedding Predictive Architecture (JEPA) and trained on a large-scale dataset spanning 12 robotic embodiments. We show that RoboJEPA's imagination error, the error of its latent rollouts, follows a second-order power law in compute, allowing us to predict model quality well beyond the scale at which the law is fit. We further show that downstream robotic planning performance improves predictably with compute, and that imagination error is strongly correlated with it, making it a reliable proxy for real-robot evaluation. Finally, we demonstrate that latent world models can be deployed zero-shot as robotic agents, planning toward a single goal image to solve tasks requiring long-horizon planning on real hardware. We release all model checkpoints together with our training and robot deployment code. To our knowledge, this is the first work to establish scaling laws for multi-embodiment robotic world models trained on real robot data, and RoboJEPA, at 8B parameters, is the largest JEPA predictor model trained to date.

0.40

Kernel Autoresearch for Open-Ended Model Discovery

cs.LG ·2026-10-07 ·Richard Cornelius Suwandi, Feng Yin, Kevin Murphy
highly cited
Abstract

Kernels encode the inductive bias of a wide range of machine learning models, yet automated kernel design faces a fundamental dilemma. A fixed grammar of base kernels and operators guarantees validity but limits the search to structures expressible by those building blocks. Conversely, unrestricted programs remove this limitation but no longer guarantee validity. In our stress tests, 22-58% of LLM-generated kernels that pass numerical checks on random inputs fail when evaluated at different scales or dimensions. We propose Kernel Autoresearch (Kernaut), which treats kernel design as open-ended model discovery. Coding agents write kernels as programs, while construction contracts ensure that every accepted kernel is valid. A quality-diversity archive retains high-performing kernels with distinct behaviors, and novelty screening steers agents toward functionally new candidates. Our experiments demonstrate that the discovered kernels encode reusable inductive biases that generalize to unseen tasks. On held-out black-box optimization families, a discovered kernel outperforms a meta-learned deep kernel trained on the same episodes. Furthermore, kernels discovered from ten enzyme-kinetic rate laws achieve lower error than tuned ARD and deep kernel baselines on five unseen mechanisms. The discovered kernels are also interpretable programs that human researchers can refine: a human-refined version of one further reduces the held-out predictive error by 5.7% and optimization regret by 7.8%.

0.40

Reconstruction of Multiscale Plasma Dynamics Across Operating Regimes

cs.LG physics.plasm-ph ·2026-10-07 ·Maryam Reza, Farbod Faraji
hot direction
Abstract · 27 pages, 17 figures

Reconstructing spatially resolved plasma dynamics from few sensors is essential for diagnostics, reduced-order modelling and control, yet remains difficult because the sparse measurements incompletely constrain multiscale, regime-dependent degrees of freedom. The Shallow Recurrent Decoder (SHRED) partially addresses spatial sparsity by using measurement histories; however, its fully connected decoder provides no explicit mechanism for resolving spatial structure across scales or explicit parametric dependency. We introduce the Recurrent Multiscale Affine-modulated Inference Network (ReMAIN), which preserves SHRED's recurrent temporal encoding but replaces its decoder with a U-Net whose feature hierarchy is conditioned by the recurrent state through feature-wise linear modulation. The temporal representation supplies both a dense prior and scale-specific modulation throughout the U-Net. A parametric extension jointly embeds the operating condition and sensor history, enabling reconstruction to adapt as the governing dynamics change with operating regime. ReMAIN is first benchmarked against SHRED on six one-dimensional nonlinear PDEs representing diverse dynamics. Across all benchmarks, it reduces reconstruction errors on unseen trajectories and more faithfully resolves sharp transitions, localized extrema and fine-scale variations. The parameter-conditioned model is then demonstrated on a collisionless $E \times B$ plasma subject to perpendicular axial electric and radial magnetic fields, with the electric-field strength serving as the operating parameter. ReMAIN reconstructs the high-dimensional, multiscale plasma state and recovers its regime-dependent spatiotemporal dynamics at electric-field strengths withheld from training. Together, ReMAIN improves sparse-sensor full-state reconstruction and, through parameter conditioning, generalizes across plasma operating regimes.

0.39

NL2Hull: A Natural Language-Driven Constrained Ship Design Decision Framework

cs.AI ·2026-10-07 ·Wenhua Huo, Fenglei Han, Wangyuan Zhao, Jialin Wu, Jiayi Han
benchmarkdatasetcode/data
Abstract · 23 pages, 8 figures, 10 tables

Ship-form design combines smooth geometric representation, local shape editing, and constraints on the resulting hull. We present the Natural-Language-to-Hull Framework (NL2Hull Framework), which formulates ship-form editing as a typed discrete decision problem and connects language decisions to numerical geometry. Its Constrained Free-Form Deformation Engine (CFFD Engine) represents hull waterlines with non-uniform rational B-splines (NURBS), applies free-form deformation (FFD) to their control points, reconstructs the hull, and checks geometric constraints. We construct the Ship Design Decision Dataset (SDD Dataset) with 134,558 cleaned records and evaluate compared models on its subset Ship Design Decision Benchmark (SDDBench), containing 5,000 records and 43,496 typed questions. We propose Chip, a constrained ship-design decision model for processing natural-language requests. Chip reaches 95.90\% question accuracy and 99.32\% FFD exact match, with a negative log-likelihood of 0.0951, an expected calibration error of 0.0032, and a Brier score of 0.0551. The NL2Hull Framework provides a reproducible interface for evaluating language-based ship-form decisions while identifying the geometry and continuous-control components that require further development. Our code and dataset is available at https://github.com/wenhuahuo/NL2Hull.

0.39

Sample-Efficiency of Kolmogorov-Arnold Networks

cs.LG ·2026-10-07 ·Kevin Riehl, Shaimaa K. El-Baklish, Fan Wu, Anastasios Kouvelas
benchmarkcode/data
Abstract

Deep reinforcement learning has achieved substantial performance gains over classical control approaches. Yet, a central challenge to learning in real-world applications is acquiring costly samples. Kolmogorov-Arnold Networks are a recently proposed architecture that can learn physical relationships in control problems effectively, with significantly higher parameter efficiency and interpretability when compared to Multi-Layer-Perceptron architectures. In this work, we systematically study sample-efficiency using computational experiments, covering the Feynman dataset and the Gymnasium RL benchmark. The results show that similar performance can be achieved with 40% fewer samples using the Kolmogorov-Arnold architecture, and that relative performance improvements up to 50% occur during the training process. The observed gains are robust to varying levels of noise in rewards. These results highlight the potential of the Kolmogorov-Arnold architectures for more sample-efficient reinforcement learning. Code: https://github.com/DerKevinRiehl/neurips26_kan_training

0.39

Global Exponential Convergence of Two-Layer Linear Network Training

cs.LG math.OCstat.ML ·2026-10-07 ·Stephen Y Zhang, Gabriel Peyré
theory
Abstract

We prove global exponential (linear) convergence with an explicit rate in the rich scaling for wide two-layer linear networks trained with smooth Polyak-Lojasiewicz predictor losses. Gradient flow in the factors closes exactly in terms of a finite-dimensional Bures flow of the neuron law covariance, in which the predictor dynamics are preconditioned by hidden covariance blocks. Mean-field conservation laws provide uniform spectral lower bounds on the hidden preconditioning blocks when the initial covariance satisfies a spectral support gap condition. This condition encompasses positive definiteness while still allowing for singular initializations. For an initial covariance $Σ_0 = σ^2 \mathrm{Id}$, the loss converges to the global minimum with linear rate at least $4σ^2κ$, where $κ$ is the PL constant. We establish stability of this rate under finite-width sampling, as well as global convergence of factor gradient descent for an explicit stepsize interval depending on smoothness, the initial loss, and conserved spectral margins. Our argument extends layerwise to deep linear ResNets, subject to a residual-path bound. In the case of heavy-ball momentum, training dynamics close instead over positions and velocities in terms of a lifted phase covariance. Linear convergence holds under an explicit condition on the energy and damping, specifying a window of admissible dampings. For two-scale white initializations, this interval is nonempty for sufficiently large position scales, with a fixed initial loss gap and velocity covariance. Numerical experiments illustrate the covariance geometry and compare the predicted and observed rates.

0.39

TopoGraphRAG-Bench: Evaluating Multimodal GraphRAG on Layout-Grounded Evidence Reasoning

cs.AI cs.CL ·2026-10-07 ·Ruochi Li, Jianzhe Lin, Haoxuan Zhang, Haihua Chen, Junhua Ding +2 more
top venuebenchmarkcode/data
Abstract · Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)

Real-world documents distribute evidence across text, tables, figures, and captions within complex page layouts. Answering complex questions over such documents therefore requires more than retrieving relevant passages: systems must recover the evidence topology that connects heterogeneous evidence units. Existing GraphRAG evaluations remain largely text-centered, while multimodal document RAG benchmarks assess cross-modal retrieval and generation without directly evaluating recovery of the intended evidence topology. We introduce TOPOGRAPHRAG-BENCH, a layout-grounded benchmark for multimodal evidence reasoning in GraphRAG, comprising 2,024 questions over 201 long, visually rich documents. Questions are constructed bottom-up from text, figure, and table evidence units under three controlled topologies: single-hop retrieval, bridge-chain reasoning, and multi-source synthesis. To ensure that questions preserve their intended structure, we apply counterfactual validation for shortcut resistance, modality necessity, and evidence necessity. We evaluate text-only GraphRAG, page-level visual retrieval, and multimodal GraphRAG systems using retrieval, generation, and topology-aware reasoning metrics. Multimodal GraphRAG systems achieve the strongest overall performance, but still fail when visual-textual evidence alignment or multi-unit composition is incomplete. Text-only GraphRAG struggles when key dependencies are grounded in figures or tables, while page-level visual retrieval lacks the fine-grained structure needed for topology recovery. These findings motivate GraphRAG systems that move beyond text-derived entity relation graphs to explicitly model document layouts, cross-modal evidence alignment, and the reasoning roles of evidence units. Code and data are available at https://richardlrc.github.io/TopoGraphRAG-Bench/.

0.39

World-Model Policy Arbiter for Goal-Conditioned Reinforcement Learning

cs.LG ·2026-10-07 ·Junwei Quan, Evgenii Opryshko, Nicholas Rhinehart, Igor Gilitschenski
code/data
Abstract · 22 pages, 3 figures, 14 tables. Project page: https://junwei0102.github.io/wmpa-webpage/ Code: https://github.com/junwei0102/World-Model-Pol…

Offline goal-conditioned reinforcement learning (GCRL) has produced a diverse set of goal-reaching algorithms, yet no single algorithm performs best across environments, goals, and even different phases of the same task. Rather than deploying only the best-performing policy, we ask whether a set of frozen goal-conditioned policies can be used collectively as a portfolio, deciding at every state which policy should act. Choosing a policy at each state is not straightforward. The policies' own value functions cannot be compared directly: they may use different scales, and some policies have no value function. We need to judge each policy by the states it is likely to reach, even though we can execute only one policy at a time. We also need to avoid switching so often that control becomes unstable. To address these challenges, we introduce World-Model Policy Arbiter (WMPA), a test-time framework that, given a set of frozen policies as input, rolls out each frozen policy in a learned state-space world model, evaluates the imagined futures with a shared goal-conditioned value function, and executes the highest-scoring policy for a short commitment interval before the next round of arbitration (policy selection). WMPA assumes access to a bank of frozen goal-conditioned policies and requires neither policy retraining nor privileged task-specific knowledge. Under the official OGBench evaluation protocol on 18 state-based datasets spanning maze navigation as well as cube, scene, and puzzle manipulation, WMPA improves the macro-average success rate from the 44% achieved by the best policy selected per dataset to 58%, with statistically significant gains on 12 datasets. These gains include +33 percentage points on cube-double-play and +36 percentage points on scene-play.

0.39

When Rank Rises as LLMs Degrade

cs.LG cs.CL ·2026-10-07 ·Zhaohui Geoffrey Wang
published
Abstract · NeurIPS 2026 Workshop on Continual Learning for Foundation Models and Agents (CL4FMAgents); 8 pages + appendix

Post-training adapts language models in non-stationary environments. Practitioners monitor representation health with RankMe and related spectral statistics, often assuming that rank falls when representations degrade. We show that this assumption is unsafe for LLM post-training. In a controlled study of Qwen3-0.6B with four degradation modes and three seeds, data duplication worsens held-out loss by 75% relative to healthy while increasing both original and centred RankMe; the latter changes by 13.5 pooled standard deviations. Covariance effective rank rises to nearly twice its healthy value. This failure is spectral dispersion rather than collapse, so a one-sided monitor rates the worst checkpoint as the healthiest. By contrast, a learning-rate misconfiguration lowers centred RankMe and k95, while uncentred RankMe is inconsistent across seeds. Direction is therefore a property of the regime-statistic pair and cannot be fixed by recalibration alone. We also distinguish two often-conflated statistics: RankMe normalises singular values, whereas covariance effective rank normalises eigenvalues. On raw intermediate-layer states in the pretrained model, massive activations pin the latter near 1 out of dimension d while RankMe retains usable range. We then test a two-sided, multichannel sequential monitor with separate calibration and test data. In a pre-registered shared-prefix, leave-one-seed-out evaluation, it detects all three damage regimes in every fold 10 to 60 steps after the fork and separates dispersion from downward-rank damage by firing direction. However, it never precedes held-out probe loss, and calibration with two seeds produces false alarms on the held-out healthy seed. Spectral monitoring can diagnose failure regimes, but it does not warn earlier than held-out loss, and validity claims require held-out healthy data.

0.39

Data Reuse in Non-Stationary Learning

cs.LG stat.ML ·2026-10-07 ·Tomer Gafni, Garud Iyengar, Assaf Zeevi
theory
Abstract

We consider online learning in non-stationary environments, where the goal is to track an unknown parameter that switches abruptly between a finite set of recurring values. Recurrence opens the possibility of judiciously reusing past observations to improve algorithm performance. However, the changing nature of the underlying signal and lack of information on these dynamics may limit the ability to "safely" reuse data. In this paper we quantify some of the fundamental tradeoffs in this class of problems, and show that they bear a certain resemblance to the classical bias-variance dilemma. Specifically, we propose a class of anytime algorithms, dubbed Exposure-Capped Reuse (ECR), that combine online change detection, compatibility testing, and "contamination" control. We characterize the regime in which ECR's regret scales with the number of distinct values rather than the number of changes, and derive a novel information-theoretic lower bound that establishes the near-minimax optimality of ECR. This provides rigorous quantification of the statistical "value" of data reuse.

0.38

RADAgent: Automating RADseq Pipelines with Private Open-Source LLMs

cs.AI q-bio.GNcs.SE ·2026-10-07 ·Qiming Guo, Malaka Sudhakara Bava Sukruth Reddy, Honggang Zhao, Ye Wang, Chen Pan +1 more
publishedcode/data
Abstract · Accepted at IEEE BIBM 2026. Code: https://github.com/wenlu-lab/RADAgent

Restriction site-associated DNA sequencing (RADseq) is a family of reduced-representation sequencing methods. It uses restriction enzymes to sample subsets of the genome, therefore allowing multiple individuals to be sequenced and genotyped together at low cost. RADseq is widely used for population genomics, evolutionary biology, marker discovery, and the development of SNP panels, but complete analysis still requires expert coordination across multiple tools and processing stages. Traditional RADseq analysis requires expert coordination of heterogeneous bioinformatics tools, intermediate artifacts, quality-control checks, and parameter-dependent decisions across multiple stages. Although individual steps can be scripted, executing the complete workflow still requires human oversight for state tracking, result validation, and failure handling. To automate this expert-driven coordination, we present RADAgent, which introduces a modular agentic method for reliable RADseq workflow execution with an open-weight LLM. RADAgent maps the natural boundaries of the RADseq workflow into 16 independently executable and verifiable agent modules and uses a model-independent runtime for bounded tool execution, on-disk verification, and failure recovery. We evaluate RADAgent on a real population-genomics dataset containing 476 samples from 19 populations. The results demonstrate that RADAgent reliably automates the complete RADseq workflow with an open-weight LLM while keeping genomic data and model execution within the local computing environment. The code is publicly available at https://github.com/wenlu-lab/RADAgent.

0.38

Constrained Diffusion for Data-Scarce Orbital Monte Carlo in Constellation Tasking

cs.AI astro-ph.IMastro-ph.EP ·2026-10-07 ·Omar Ramadan, Sam Siavoshian, Amir Kashif Saeed, Benjamin A. Johnson, Amin M. E. -A. Diab +1 more
Abstract · 16 pages. Submitted to the 2027 IEEE Aerospace Conference; under review

Constellation Monte Carlo results depend on the orbital population used to evaluate a tasking policy. With scarce reference trajectories, replay limits geometric diversity, while independent orbital-element jitter can violate physical constraints. We study constrained diffusion for orbital-population augmentation. A force-conditioned diffusion model learns a 13-dimensional orbital prior, recovering semimajor axis from perigee altitude and eccentricity; Basilisk propagates each sample under one of five force-model tiers. Using 800 reference trajectories, we compare diffusion with jittered bootstrap, per-tier Gaussian mixtures, and a conditional variational autoencoder, and evaluate distributional fidelity, support shift, classical astrodynamics diagnostics, and 4,000 paired GoDSAT-compatible campaigns. The largest diffusion model achieves held-out trajectory MMD of 0.0171 +/- 0.0239, similar to bootstrap (0.0170) and the mixture (0.0198), but samples farther from training priors (median nearest-training distance 2.7 versus 0.10 standardized units). All generators fail to match the shifted-blind population (classifier AUC 0.991-0.998). Endpoint-conditioned samples satisfy Lambert boundaries but have greater interior error than the matched Lambert reference (29.2 versus 5.7 km mean RMSE). Residual diffusion improves selected sparse forecasts and catalog-mode recall but does not outperform classical estimators on custody ranking. In a fixed 16-satellite configuration, diffusion yields similar mean custody to replay and bootstrap, while the force-tier mixture shifts custody by about 5.5 percentage points. Constrained diffusion supports local, in-support augmentation but cannot replace orbital dynamics or serve as an operational posterior. Orbital-population construction is a consequential source of uncertainty in constellation analysis.

0.38

When Routing Reveals Membership: Privacy Leakage from MoE Router Telemetry

cs.LG cs.CR ·2026-10-07 ·Yixin Tan, Jiayang Liu, Lu Sun, Yuke Hu, Zheng Li +1 more
Abstract

Mixture-of-Experts (MoE) language models produce routing information during inference that may be logged or exposed for monitoring, debugging, load analysis, and safety auditing. Unlike ordinary model outputs, this telemetry reveals a view of the model's internal computation, raising a privacy question: can it reveal whether an example was used to fine-tune the deployed model? We introduce a router-augmented membership inference attack that combines conventional output-side signals with aggregated routing features and applies a membership classifier learned from independently fine-tuned shadow models to the target model. Across three MoE architectures and three data domains, router telemetry consistently improves membership inference over a strong output-signal ensemble, increasing TPR at 1\% FPR by 2.7--9.4 percentage points across all nine settings. The leakage persists across full fine-tuning, frozen-router training, LoRA, and instruction tuning, and remains observable with only discrete expert selections, restricted telemetry, or a single shadow model. Mechanistic analysis further shows that the leakage does not require router-specific memorization: fine-tuning introduces membership information into hidden representations, while the router exposes a projection of this signal even when its parameters are frozen. Perturbing the telemetry reduces this additional leakage only as its fidelity degrades. Our results show that router telemetry can turn an operational signal into an additional privacy surface for fine-tuned MoE models.

0.38

Possibilistic Radial Transport for Approximate IM Inference

cs.LG stat.ML ·2026-10-07 ·Jungeum Kim, Percy Zhai
Abstract

Probing the hypothesis space after seeing the data remains valid under possibilistic inferential models (IMs), provided the significance level stays fixed. The price is computation, as each plausibility is a supremum of the possibility contour over the hypothesis, and the contour itself is approximated at each queried parameter value. We propose a possibilistic radial transport, which hides the contour value of a parameter in the radius of its source point. When a transport that maximizes within-shell entropy is picked, sampling parameters covering a confidence cut becomes a matter of truncating the radius. We provide a deep learning algorithm that enforces the contour depth condition while maximizing the entropy within each shell. Our amortization makes coverage and power assessments of the learned approximation practical as well as predictive check of new datasets. We also use the sampler to construct a Bel-Pl spectrum for comparing and selecting interpretable hypotheses that satisfy a prescribed Bel-Pl decision criterion. In simulations the learned contours match or improve on ellipsoidal approximations to the cuts, while the coverage and power track the exact reference. Finally, we probe hypotheses about ovarian aging using synthetic AMH records, asking for each woman how many more years her median AMH level will remain above a specified reference value.

0.38

Fully Interpretable Minimal Transformers: From Geometry to Algorithm

cs.LG cs.AI ·2026-10-07 ·Raneem Mahajne, Toviah Moldwin
code/data
Abstract · 27 pages, 15 figures, 2 tables. Code and training-dynamics animations: https://github.com/Raneem-mahajne/creating_transformer

We present a framework for building and interpreting minimal transformer models. By constraining a transformer's embedding dimension and head size to 2, we enable full two-dimensional visualization of its internal representations. Embeddings, query/key/value transforms, attention outputs, residual streams, and decision boundaries can all be seen directly. Our central claim is that the learned geometry implies an algorithm; the arrangement of points and boundaries in R^2 can be read as a step-by-step procedure. We train a transformer on a simple task where it must produce the most recently observed even number whenever the '+' operator appears in a sequence of digits. Once trained, we visually walk through every step of the transformer's computation. We show how the model embeds the tokens and their respective positions in the sequence, transforms them via the Q, K, and V matrices, uses the dot product between the Q and K representations to form the attention matrix, and uses the attention matrix to select values that move the representation of each input token to the region of the domain of the output layer that will correctly predict the next token. We introduce a suite of interpretability visualizations that make the algorithmic interpretation of this procedure explicit. Our framework offers a pedagogical and experimental testbed to explore how transformers use informational geometry to implement next-token prediction.

0.38

EdgeRock: Strong Reasoning at Constant Memory

cs.LG cs.CL ·2026-10-07 ·Huishan Ji, Weiming Zhang, Qirui Ye, Jinchao Xu, Yike Guo +1 more
code/data
Abstract · 24 pages, 8 figures. Substantially revised technical report under the EdgeRock name, with expanded evaluation and deployment results. Code: …

We present EdgeRock, a family of multimodal models that deliver strong reasoning at constant memory. Most models keep reasoning history in a key-value (KV) cache that grows with every token, consuming accelerator memory, limiting concurrent requests, and competing with perception and control on edge devices. With fixed history, memory is allocated once and never runs out, however long the model reasons. Constant memory alone is not new: bounded-state models (recurrent, state-space, and sliding-window) already achieve it, but often at the cost of reasoning, because they cannot keep the distant details that later reasoning needs. We therefore introduce Rock Memory Anchors (RMA): the model learns what is worth remembering and stores it as KV records in a fixed-size bank, directly readable however far back the source lies. EdgeRock-4B, our first open-source release, supports text and image reasoning and reaches 82.9% mean pass@1 on AIME 2024-2026, while the bounded-state models we evaluate reach at most 18.8%, even at 14B scale. This result is on par with leading compact models whose memory grows with reasoning. Compared with its source model, EdgeRock-4B uses up to 99.1% less KV cache and delivers up to 6.42$\times$ higher end-to-end batch throughput.

0.38

Homogenization in Multi-Agent Systems

cs.LG ·2026-10-07 ·Prakhar Ganesh, Kyra Wilson, Luca Zappella, Barry-John Theobald, Nicholas Apostoloff +2 more
hot direction
Abstract

Multi-agent systems (MAS) leverage interactions between agents to perform complex tasks. Despite their success, we show that these interactions can also lead to homogenization, i.e., agents converging to similar behaviors. Homogenization in MAS can reduce agent diversity and reinforce shared failures. In this paper, we operationalize homogenization using three metrics: conformity to the majority, polarization towards extremes, and growing inertia against changes over subsequent interactions. We evaluate homogenization in MAS for code generation, hiring, and scientific peer review. Across these tasks, we show that homogenization translates to concrete downstream risks: in code generation, it hides and amplifies correlated errors which can create systemic vulnerabilities; in hiring, it allows the influence of biased agents to persist long after their removal; and in peer review, it creates uneven evaluation standards across research areas. Our results establish homogenization as a failure mode of MAS, demonstrating that MAS evaluations must move beyond aggregate performance to carefully analyze interaction dynamics. Finally, we show that simple approaches to increase diversity---leveraging sampling stochasticity and mixed-models MAS---fail to reduce homogenization risks, highlighting the need for strategies to effectively leverage agent diversity.

0.38

Auditing LLM Program Search for Bermudan Stopping with Primal--Dual Certificates

cs.LG cs.AI ·2026-10-07 ·Yuhua Qian
publishedtheory
Abstract · Accepted at the MATH-AI Workshop at NeurIPS 2026

LLM program-search agents are usually judged by the score that drives their search. We study LLM-driven search over Python exercise policies for Bermudan options, an optimal stopping problem where a trusted evaluator owns the simulation and a dual upper bound~$U$. For any policy, $0 \le V^\star - L_π\le U - L_π$; we report Monte Carlo estimates of both bounds with their sampling error. We audit 24 generation attempts from two pipelines, a six-pair feedback-representation ablation, and a holdout evaluated after selection was frozen. No selected generated program improves on Ridge LSMC consistently or by more than Monte Carlo uncertainty; holdout gap scores lie within $0.002$ of it. On six one-dimensional puts, a high-precision binomial reference places Ridge within Monte Carlo noise of optimal and attributes 99\% of the estimated gap to dual-bound looseness, so this dual audits validity but cannot certify progress at the search's scale. Structured memory showed no benefit over score-only memory in this ablation. Whether search ``worked'' depends on reporting rules: retaining the initial reference, freezing selection before holdout, and counting failures.

0.37

Seq-Flow: Efficient Probabilistic Forecasting with Self-Rollout Error Control

cs.LG ·2026-10-07 ·Yinan Huang, Shitij Govil, Bo Dai, Pan Li
code/data
Abstract

Many scientific forecasting tasks require updating a distribution over future trajectories as new observations arrive. Conventional diffusion and flow models generate each forecast from Gaussian noise, often at the cost of many sampling steps. Warm-start methods reuse earlier predictions to reduce this cost, but their models are not trained to perform the forecast update itself, which can compromise quality under few-step sampling. In this work, we introduce Seq-Flow, a conditional flow model whose ODE transports samples from the previous forecast distribution to the updated one. Because successive forecasts often differ only modestly, this transport starts from an informative distribution and can produce accurate updates with few flow evaluations. Recursive reuse also creates a challenge: errors in one forecast become errors in the initial states of subsequent flows. We address this with self-rollout training, in which a moving average copy of the model generates forecasts that initialize later training updates. Unlike self-forcing methods, which reuse generated outputs as conditioning context, Seq-Flow reuses them as the source of the next flow. Experiments On particle-accelerator beam spill forecasting show Seq-Flow reduces CRPS by 65% under a few-NFE sampling budget, while remaining competitive with strong baselines on fluid-dynamics forecasting tasks. Although trained on self-rollouts of at most four updates, Seq-Flow remains accurate over more than 400 consecutive updates. Our code is available at https://github.com/Graph-COM/Seq-Flow.

0.37

Reasoning-Token Spikes Under Prompted Untruthful Responses in Large Language Models

cs.AI cs.CL ·2026-10-07 ·Maverick Morales, Tomáš Dominik, Vermut Gao, Katrina Shirey, Paulius Rimkevičius +2 more
code/data
Abstract · 21 pages, 9 figures, 3 tables. Code: https://github.com/Wakaranaino/token-spike-project ; Data: https://doi.org/10.5281/zenodo.21895296

Monitoring the chain-of-thought of reasoning artificial intelligence (AI) models remains a key approach to detecting deception and other forms of misbehavior in such models. However, semantic chain-of-thought monitoring depends on reasoning traces being legible and sufficiently faithful to the underlying computations that produced the model's behavior, not to mention accessible. Moreover, there is increasing evidence that chain-of-thought outputs may soon become illegible or unfaithful, if they even remain accessible. Based on cognitive load theory, we investigate a lower-bandwidth signal--the number of reasoning tokens generated--which does not require access to the content of the reasoning trace. Three reasoning-capable large language models answered 210 multiple-choice questions--across analytic, descriptive, and normative reasoning types as well as moral and non-moral domains--under system prompts instructing them to respond truthfully, falsely, or without regard for truth. Across all three models, truth-directed responses elicited fewer reasoning tokens than both lie-directed and truth-indifferent responses. In addition, in one model, untruthful instructions increased the likelihood of reasoning at all. These findings show that explicitly prompted untruthful response policies can produce robust group-level differences in test-time reasoning-token use. While not yet establishing reasoning-token count as a detector of spontaneous deception or general misalignment, our results are a proof of concept that it can serve as a simple, content-independent black-box candidate signal for differentiating untruthful from truthful model behavior when raw reasoning traces are unavailable or unreliable. Future work should test instance-level detection rates, out-of-distribution generalization, learned deceptive policies, hidden objectives, and robustness under adversarial pressure.

0.37

How Many Directions Must a Truncated Diffusion Sampler Retain? Matching Bounds Under Power-Law Spectra

cs.LG math.STstat.ML ·2026-10-07 ·Radmehr Karimian, Ali Mohades, Johannes Lederer
theory
Abstract

Diffusion samplers can reduce computation by generating selected spectral coordinates and filling the remaining directions with noise. How many directions must they retain? We study this question for data with power-law covariance spectra. For Gaussian data compared to a smoothed target, we prove matching bounds on the required number of retained directions, provided that the ambient dimension is sufficiently large. The truncation error depends on the combined Wiener gains of the omitted directions, regardless of the accuracy of the sampler on the retained coordinates. Keeping only directions whose signal exceeds the output noise level can therefore leave a non-vanishing error: many individually weak directions remain significant in aggregate. Combining this characterization with a diffusion convergence bound yields sufficient sampling-step complexity under exact scores. The upper bounds also extend to estimated principal components and, componentwise, to Gaussian mixtures. The practical prescription is to select the retained subspace using an aggregate spectral-tail error budget, then to choose the diffusion noise level accordingly.

0.37

CHASE: Channel-Aligned Structure Exploitation for Geometry-Aware Model Engineering

cs.LG cs.CL ·2026-10-07 ·Wei Wang, Wei Jiang, Ziran Liu
Abstract · 26 pages, 9 tables

Geometric and Spectral Alignment (GSA) characterizes trained networks through spectral concentration, physical-channel alignment, support structure, and changes in singular bases. In this paper, we propose CHASE (Channel-Aligned Structure Exploitation) to use these structures in practical model design. CHASE covers six applications across model modification, reconfiguration, and compression. CORA, COEC, and CORAM apply GSA to parameter-efficient finetuning, structured-pruning compensation, and model merging. We further develop three new methods. CAGA uses GSA to identify multi-head attention heads that can share a KV representation and constructs the shared key and value heads through geometric alignment and low-rank subspace extraction. SAKV uses GSA to determine which adjacent layers can share a low-rank KV-cache representation and the retained rank for each layer group. CAPS uses GSA spectral structure to group output neurons and selects retained input channels separately for each group. Results from CORA, COEC, and CORAM establish the effectiveness of GSA for adaptation, pruning compensation, and model merging. Experiments on CAGA show that geometric shared-head construction substantially improves MHA-to-GQA conversion, and SAKV and CAPS improve over representative baselines for KV-cache compression and structured pruning. These results show that the structures identified by GSA can be used directly to design methods for a range of model operations.

0.37

Input-Blind Controls Produce Substantial Oracle Headroom for Layer Programs in Multiple-Choice Evaluation

cs.LG cs.CLcs.SE ·2026-10-07 ·Yibei Guo, Rui Liu
Abstract

Adaptive computation aims to improve language-model inference by tailoring execution to each input. For layer programs, oracle evaluations use known answers to estimate the potential gain from this flexibility, before a practical selector is available. However, a gain from selection does not by itself explain why the chosen programs help. This study examines this distinction using 32 layer-skipping and repetition programs on two models and 4,413 multiple-choice items. The analysis compares their gains over a fixed action selected without the evaluation prompt with those of input-blind perturbations at the same sites, re-evaluating selections on another prompt. With shared option order, the controls give 10.2-11.8 and 15.6-19.4 percentage points of headroom on Qwen3-4B-Base and Llama-3.1-8B, exceeding the real programs' 9.0 and 10.1 in all three random-direction draws per model. They match answer-change rate only, and the ordering depends on the menu: in post hoc comparisons, real programs lead on Llama's repeat-only menu in every draw. A smaller KL-calibrated comparison, including an input-dependent control, favours real programs in point estimate, with inconclusive corrected tests. Fixed letter offsets produce headroom of similar scale. Rotating options sharply reduces both families' headroom, while leaving positive real-minus-control differences of 1.4-2.3 and 3.7-4.5 points; their magnitudes and statistical support depend on further adjustments and the reference. A supplementary generated-answer test finds that search-selected programs keep a 26.0-point advantage over programs selected for other problems after rewording, without a placebo comparison. These results show that substantial headroom can persist across prompts with shared option order without establishing a benefit specific to the selected layer computation; neither ordering against these controls identifies that benefit.

0.36

Collaborative Reasoning Distillation via Cross-Feedback and Coherent Curation

cs.LG cs.AIcs.CL ·2026-10-07 ·Taehoon Kim, Seunggeun Cho, Dongsu Han
top venue
Abstract · Accepted at NeurIPS 2026

Reasoning capabilities are critical for advancing Large Language Models, yet current approaches either require massive computational budgets or struggle to effectively distill reasoning to smaller models. Standard distillation methods rely on outcome-based rewards, failing to distinguish between sound reasoning and lucky guesses. We propose Collaborative Reasoning Distillation (CRD), a framework that enhances reasoning in compact models through three innovations: (1) interactive cross-feedback where teachers iteratively critique each other's reasoning, (2) fine-grained step-wise quality assessment capturing logical validity independent of final answers, and (3) coherence-aware step stitching that synthesizes complementary strengths. Students are trained via Reasoning Quality Optimization (RQO) with budget constraints. Our model, CRD-4B, achieves 97.3% on MATH-500 and 70.3% on AIME'25, surpassing baselines while using only 50K training examples, up to 12 times smaller than the datasets of comparable models.

0.36

Training Parallel Speculative Draft Models by Directly Minimizing Expected Decoding Rounds

cs.LG cs.AIcs.CLstat.ML ·2026-10-07 ·Yunxiao Zhao, Changxiao Cai
Abstract

Speculative decoding accelerates large language model inference by using a low-cost draft model to propose tokens that the full-size target model verifies in parallel. Parallel and semi-autoregressive (semi- AR) drafters improve drafting efficiency by proposing an entire block in a single forward pass, but training them raises a new difficulty: the draft distribution for a given position depends on where the decoding round starts, and where rounds start depends on how many tokens earlier rounds accepted. Existing training objectives typically rely on block-local surrogates that ignore this cross-round coupling, and therefore do not directly optimize the global decoding efficiency. In this work, we develop a theoretical framework for training and evaluating these drafters by representing speculative decoding as a Markov reward process. This formulation yields the Expected Decoding Rounds (EDR) objective, which weights local rejection costs by state occupancies and exactly equals the expected number of decoding rounds. Unlike prior surrogate objectives, EDR introduces no auxiliary hyperparameters. We then derive an exact temporal-difference gradient that supports unbiased stochastic optimization from target-model rollouts. The same framework also yields an exact offline evaluator for round counts, enabling paired drafter comparisons on shared target rollouts without running speculative decoding. Finetuning two state-of-the- art drafters, DSpark and DFly, with EDR consistently improves mean accepted length and outperforms existing training objectives across nine benchmarks spanning math reasoning, code generation, and chat.

0.36

Safe Meta-Policy Design with Risk Control

cs.LG stat.MLmath.OC ·2026-10-07 ·Wenbin Zhou, Michael Lingzhi Li, Shixiang Zhu
hot direction
Abstract

Models can be retrained as new data arrive, but deploying every new version risks replacing a good policy with a worse one. We study how to plan policy updates (i.e., meta-policy) before future candidates are trained, balancing the benefits of improvement against the risk of performance regression. Our offline meta-policy maximizes expected cumulative value subject to a budget on the expected number of updates that perform worse than the policies they replace. We estimate the value and risk of possible switches from historical learning trajectories, represent an update schedule as a path in a directed acyclic graph, and select a schedule using dynamic programming. A leading-order analysis identifies the signal-to-noise ratio of policy improvement as a key driver of update frequency, waiting times, and risk allocation: clearer improvements support earlier, more frequent updates, while noisier improvements call for longer waits or greater risk expenditure. Their asymptotic rates also reveal a diminishing marginal cost of achieving greater safety over time. Experiments on synthetic and clinical trial data illustrate the performance--risk tradeoff and compare our method with alternative baselines.

0.36

Q-PhotoMarket: A Design Space Exploration Framework for Photonic Hybrid Quantum Neural Networks in Financial Market Prediction

cs.LG quant-ph ·2026-10-07 ·Alberto Marchisio, Hanzalah Mohamed Siraj, Muhammad Kashif, Nouhaila Innan, Muhammad Shafique
top venue
Abstract · To appear at the IEEE International Conference on Quantum Artificial Intelligence (QAI), Nottingham, UK, December 2026

Photonic quantum computing has recently emerged as a promising platform for hybrid quantum machine learning due to its native realization of linear-optical circuits and the computational complexity of boson sampling. However, despite growing interest in quantum methods for finance, the influence of photonic circuit design choices on predictive performance remains largely unexplored. Existing studies typically evaluate a single architecture, leaving the broader photonic design space unexamined. In this work, we present Q-PhotoMarket, a systematic design space exploration (DSE) framework for photonic hybrid quantum neural networks (HQNNs) applied to financial market prediction. We explore over 5,000 valid photonic configurations spanning input photon states, circuit architectures, entangling models, and measurement strategies across their compatible computation spaces, for U.S., Indian, and cryptocurrency markets. To improve search efficiency, the exhaustive exploration is complemented with Bayesian optimization. We further incorporate threshold calibration and prediction-collapse diagnostics to enable reliable evaluation under increasingly imbalanced return thresholds. Experimental results show that systematic exploration of more than 5,000 photonic HQNN configurations reveals consistent architectural patterns across financial markets, identifies robust high-performing designs, and demonstrates competitive performance relative to classical machine learning baselines.

0.36

OOM-RL II: Reality Is an Oracle, Not a Debugger Provenance-Constrained Diagnosis in Continually Evolving Agent-Engineered Systems

cs.AI cs.SEq-fin.PM ·2026-10-07 ·Kun Liu, Liqun Chen
Abstract · 38 pages, 14 figures, 9 tables. Supplementary Dataset S1: https://doi.org/10.5281/zenodo.23215521. Follow-up to arXiv:2604.11477

Reality may establish that an outcome occurred without identifying which evolving procedure produced it or why. This distinction matters in production ML systems whose code, configuration, and artifacts change while external feedback accumulates. We examine it in a human-directed, agent-engineered quantitative trading system, using oracle to mean an external source of realized outcomes rather than a complete correctness specification. Across one year, the account gained and outperformed a broad market index, while annual alpha was not statistically distinguishable from zero under the main retrospective specification. Retrospectively selected subperiods include adverse relative performance and conditional candidate-level weakness under declared approximate references. Engineering records document changes during the episode, and complete recommendation-to-runtime binding is unavailable. The archive does not establish a common frozen instance or a unique cause. The case motivates an outcome--diagnosis gap: outcome evidence, evaluated-object identity, and causal explanation support distinct claims. We distinguish frozen instances, pre-specified adaptive procedures, and ad-hoc development; organize archive-relative claim identifiability and an evidence hierarchy; and propose a prospective production-binding protocol. An illustrative compatible-history example shows how factual binding can resolve a recommendation's referent without supplying its counterfactual effect. The protocol is proposed rather than prospectively validated. External feedback constrains outcome claims, while provenance and additional identification structure determine the resolution of diagnosis.

0.36

Pretraining Shapes Spectral Structure: Architecture- and Strategy-Conditional Prediction of OOD Robustness in Foundation Models

cs.LG ·2026-10-07 ·Sangyoon Bae, Sk Miraj Ahmed, Shinjae Yoo, Jiook Cha
theorycode/data
Abstract · 10 pages, 3 figures

Can we determine whether a foundation model will generalize out-of-distribution (OOD) before any target data is available? Existing diagnostics require source or target data, which rules them out before a target domain exists. Those that use the weights alone apply one statistic to every architecture, and do not separate robust models from fragile ones. We show the answer is encoded in the spectral structure of pretrained weights. Two forces shape that structure. Architecture determines how information is stored in weight matrices. Pretraining strategy determines what is rewarded. Together they set a spectral geometry that governs OOD robustness. We prove that the OOD accuracy gap is bounded by how tightly the source representations concentrate. A statistic computed from the pretrained weights alone serves as a proxy for that concentration. The direction of that proxy reverses between architecture families. We operationalize it: the direction is stable within one (architecture X strategy) combination, the finest grouping we test, which we call a cell. Pooled over 116 models spanning 7 modalities, a single statistic ranks OOD robustness weakly, because cells of opposite direction cancel. Within a cell, the statistic selected for it orders 92% of model pairs by OOD robustness in-sample. The selection does not leak the target: for each model family outside the matrix we logged the cell, metric and sign before running its OOD evaluation, and the predicted direction held in every case: EEG, genomic and protein. Acting on spectral concentration narrows the OOD gap by 24% at 87.5% ID retention. The diagnostic operates on released weights alone, so OOD robustness becomes checkable at model-selection time, before data or compute is committed to a target domain.

0.36

PoreML: A Data-Driven Framework for Learning Multiphase Flow in Porous Media

cs.LG physics.flu-dyn ·2026-10-07 ·Chunyang Wang, Mingrui Zhang, Yuyan Zhang, Linqi Zhu, Xin Ju +3 more
code/data
Abstract

Multiphase flow in porous microstructures is central to CO$_2$ storage, fuel-cell operation, and flip-chip packaging. Predicting these flows remains challenging because wettability and complex pore geometry govern the nonlinear evolution of fluid interfaces. Machine learning holds substantial promise for advancing the field, but progress is constrained by scarce time-resolved 3D datasets and a lack of a unified workflow for training and evaluating models. To fill this critical gap, we introduce PoreML, an open-source framework unifying data generation, model training, and evaluation grounded in pore-scale physics. The framework comprises three core components. (a) A modern GPU-native lattice Boltzmann solver, validated against analytical solutions and published experiments, enables reproducible data generation. (b) A 3.3 TB dataset contains 560 simulation runs and 158,546 stored time steps across four application-driven scenarios. These trajectories span synthetic structures and geometries derived from micro-CT scans of real materials, covering diverse wetting conditions and viscosity ratios. (c) A unified learning framework evaluates one-step prediction and autoregressive rollouts. Its domain-specific evaluation protocols assess predictive accuracy and physical consistency. We evaluate five models of diverse architecture under these protocols. Two complementary challenges assess transfer to larger domains and from synthetic to micro-CT-derived structures. PoreML provides a shared foundation for machine-learning research on multiphase flow in porous media, with the aim of empowering the community to develop reliable predictive models and advance the field.

0.35

Organizing LLM Inference for Aggregate Predictions from Behavioral Data

cs.LG cs.AIecon.EM ·2026-10-07 ·Varun Gupta, Vijay Kamble
benchmark
Abstract

When using a large language model (LLM) to simulate a user population, how should we combine individual behavioral histories to predict the fraction of users who take a particular action? We study this question in e-commerce, predicting observed product-level conversion rates from browsing sessions without task-specific model training. We vary two design choices, each involving a bias-variance tradeoff. Summarizing a product's sessions before inference reduces cost but can discard predictive detail, adding bias, whereas predicting outcomes for individual sessions and averaging afterward preserves that detail but requires many calls to reduce variance. Evaluating products together in one prompt can help the model infer what is typical in the setting, reducing bias, but also makes each prediction depend on which other products appear alongside it, adding noise. We evaluate both choices across five LLMs and two datasets. Our main finding is that predictions from individual sessions outperform predictions from summaries when products are evaluated separately, but summaries lead when products are evaluated together. Comparative summaries achieve the highest Spearman correlation with observed rates in 12 of 15 settings, with no alternative significantly better. Comparison thus helps summaries far more than session predictions, overcoming the advantage of behavioral detail. We interpret this difference through a Bayesian model of the second tradeoff, which suggests that comparison helps more when the inputs in a prompt reveal more about differences across products, consistent with our results for four types of comparative context. Comparative summaries also incur a small fraction of the inference cost of the session methods. With the best model for each outcome, they attain 66-89% of the correlation of the best supervised benchmark trained on thousands of labeled products.

0.35

Conversational Task Disambiguation over Tabular Data: Leakage-Aware Formulation, Benchmark Suite, and Training

cs.LG cs.AIcs.CL ·2026-10-07 ·Nafiseh Ghoroghchian, Luis Scoccola, Tina Sedaghat, Omid Vaheb, Hannah Chen +2 more
benchmark
Abstract · 39 pages (9 main, 30 appendix), 12 figures (5 main, 7 appendix)

Conversational task disambiguation over tabular data uses dialogue to resolve missing information about a user's intended task before producing a solution over tables or databases. Existing evaluation and training lack a leakage-aware foundation. Task success mixes the agent's disambiguation and solution-generation capabilities and can also reflect oracle leakage, that is, information that a user simulator reveals beyond what a real user would. Existing datasets also lack a shared representation of ambiguities and access boundaries. We introduce the notion of an ambiguous verifiable task, which formalizes ambiguities and resolutions, decomposing the agent into an asking policy and a solution policy, and the environment into an oracle and verifier. This framework provides baselines and metrics for evaluating task disambiguation separately from solution generation, formal definitions of oracle leakage, judge-free leakage diagnostics, and a training objective for the asking policy. We instantiate the framework in text-to-SQL with AmbiTab, a benchmark suite that unifies six ambiguous datasets under a common representation specifying what the agent, oracle, and verifier may access. We evaluate clarification strategies and oracle leakage, and train an asking policy with reinforcement learning. The trained asker improves our disambiguation metrics on all six datasets and task success on five, and our leakage diagnostics measure how training affects oracle leakage.

0.35

Multi-Agent Coordination via Support-Preserving Distillation

cs.LG cs.MAcs.RO ·2026-10-07 ·Sangmin Lee, Youngju Na, Chanmi Lee, Sung-eui Yoon
publishedcode/data
Abstract · NeurIPS 2026, Project page: https://alex6095.github.io/mosdot/

Offline MARL increasingly relies on generative policies to model multimodal joint behavior, typically by distilling a centralized teacher into decentralized one-step actors under the CTDE. We identify a failure mode at the teacher training stage: standard flow-based teachers pair noise with replay targets independently, so nearby noise samples can be routed toward conflicting coordination modes. The teacher then produces samples between valid modes, and because the distillation loss regresses each local actor onto the conditional mean of the teacher's output given local input, this error is not absorbed but propagated to the student. To remove this teacher-side artifact, we propose Mode-Support Semi-Discrete Optimal Transport (MoSDOT), which summarizes multimodal replay into a finite mode support with prescribed capacities and uses conditional semi-discrete optimal transport to assign each noise sample to a single mode before teacher training. We additionally study a shared-randomness variant that uses a shared noise component at execution to expose the residual gap intrinsic to strict-product execution. On controlled diagnostics and offline MARL benchmarks, MoSDOT improves endpoint quality and routing consistency, particularly on datasets exhibiting multimodal joint behavior.

0.35

How Hackable Is Your Speech Quality Metric? A Corrected Protocol, a Benchmark, and What Patching Buys

cs.LG ·2026-10-07 ·Ali Alavi, Donald S. Williamson
benchmarkcode/data
Abstract

Speech quality predictors are increasingly used as rewards, yet no agreed measure of their hackability exists. The usual measurement has two flaws. First, the perturbation reaches the predictor through a processing chain -- here a neural codec -- that shifts the score on its own, which scoring against the raw input charges to the attack. Referencing the unperturbed round trip instead changes measured hackability by up to a factor of four (0.31 to 0.08 for one defence). Second, one trained attacker is a sample, not a measurement: five attackers differing only in random seed reach success rates from 0.00 to 0.38 against one fixed predictor, so a defence claim needs the worst case over several. Under this protocol, four published predictors differ widely: NISQA is hacked on 90% of utterances, SSL-MOS on 21%, DNSMOS on 14% and UTMOS on 6%. We then audit a closed attack-detect-patch loop. It hardens the predictor only in its own attack space, by less than the spread between attackers; a random-perturbation baseline matches it; and it costs up to 0.30 system SRCC out of domain. Enhancers post-trained against patched predictors hack them far less (PESQ -0.03 versus -0.23). Code, preregistration and run outputs are released.

0.35

Fault-tolerant foundation models

cs.LG cs.AIcs.AR ·2026-10-07 ·Trevor McCourt, Ila R. Fiete, Isaac L. Chuang
hot direction
Abstract

Emerging computer hardware often trades reliability for energy efficiency; here we show that large-language models (LLMs) can be trained to tolerate this unreliability, and that rather than degrading, their error resilience actually increases as they grow. Modified neural scaling laws inferred from 40,000 GPU-hours of training runs on simulated faulty digital hardware quantify this trend and suggest that models learn to compute within "good" error-correcting codes, whose relative overhead remains finite no matter how large the model gets. This finding leads us to conjecture that appropriately trained LLMs may be formally fault-tolerant; if true, running AI inference on low energy, faulty hardware may be a path to substantial energy savings over the status quo.

0.34

LeCuration: A Tiny World Model as a Data Curation Multi-Tool

cs.LG cs.AIcs.CVcs.RO ·2026-10-07 ·Mayank Sengupta, Nirmit Desai, Eric Song, Kunal Sawarkar
Abstract · 9 pages

Many applications of physical AI run within finite or closed physical worlds with a limited set of physical laws governing object behavior. Examples include robots working in a warehouse and agents moving around in a video game. In order to better organize, filter, and curate data for physical AI applications, we propose a new approach centered on the unique settings and physical laws of individual datasets. We train LeCuration, a small world model intended to serve as a data curation tool for a separate, larger downstream model. To build this model, we choose LeWorldModel (LeWM)as our latent encoder and predictor, adding a diffusion transformer (DiT) decoder to add visuals to autoregressive gameplay rollout. We find that the embeddings of this model can be used as an anomaly detection signal and as a content-based clustering heuristic, and that auto-regressively predicting the game state with this model allows us to qualitatively check for action-state consistency. This paper presents a qualitative, proof-of-concept case study on CS:GO gameplay data; we do not yet report quantitative curation metrics or downstream training results, which we identify as the key next step.

0.34

A Good Self-Teacher Meets the Student Where They Are: Joint On-Policy Learning and Teaching

cs.LG cs.AI ·2026-10-07 ·Randy Ardywibowo, Arnav Dalal, Jiantao Jiao
Abstract

Reinforcement Learning (RL) from outcome rewards suffers from sparse supervision, particularly on difficult, long-horizon tasks where successful trajectories are rare and costly to generate. On-Policy Distillation (OPD) offers an attractive alternative by providing dense token-level supervision from a stronger teacher along the student's own generations. Self-distillation methods further remove the need for a separate teacher model by conditioning the same policy on privileged information to serve as its own teacher. However, privileged conditioning alone does not guarantee that the resulting distillation update improves the student. Indeed, privileged information can lead the teacher to solve tasks through shortcuts unavailable to the student, producing supervision poorly matched to the student's current behavior. Consequently, even a higher-performing teacher can provide guidance that degrades student performance. To address this, we analyze how the choice of privileged teacher affects the student's update. We derive a necessary and sufficient condition for the teacher's local distillation update to be a positive multiple of the student's reward gradient. Our analysis suggests that the teacher should not only perform well on the task, but also provide guidance suited to the student's current capabilities. This characterization motivates a practical teacher-training surrogate that combines outcome rewards with token-level Kullback-Leibler (KL) regularization toward the student. Based on this result, we propose Joint On-Policy Learning and Teaching (JOLT), which jointly trains a single policy in two roles: a privileged teacher using a KL-regularized objective, and an unprivileged student using dense on-policy distillation. Across mathematical reasoning, coding, tool use, and terminal use, JOLT improves training efficiency and performance, with further gains from student rewards.

0.34

Exact SO(3)-Equivariant Isotropic Kernels for Rotation-Robust Neural Dynamics

cs.LG cs.CEstat.ML ·2026-10-07 ·Ridham Patel
published
Abstract · Accepted at NeurIPS 2026 Workshop NeurReps (Proceedings Track)

Neural surrogates for vector-valued partial differential equations can fit training data yet change their predictions when the same physical state is expressed in a rotated coordinate frame. We study this failure on three-dimensional Navier--Stokes dynamics observed at irregularly placed points. We introduce the Invariant-Conditioned Isotropic Kernel Neural Operator (IKNO), a compact graph model that builds local interactions from scalar quantities unchanged by rotation and vector directions that rotate with the data. Consequently, rotating the positions and velocities rotates the predicted velocity change in exactly the same way. On a held-out test set fixed after model design, training unconstrained graph models on randomly rotated examples reduces but does not eliminate their coordinate dependence. In contrast, IKNO is consistent to numerical precision, matches the forecasting accuracy of a general rotation-aware Tensor Field Network with $5.6$ times fewer parameters, and outperforms a parameter-matched graph simulator. These results show that a compact, PDE-specialized model can remove coordinate dependence without sacrificing forecasting accuracy.

0.34

Align Before You Combine: Reference Space Calibration for Supervision Without Ground Truth

cs.LG cs.AI ·2026-10-07 ·Jackson Eshbaugh, Jorge Silveyra
benchmarkcode/data
Abstract · 25 pages; 14 tables; 4 figures; 5 appendices; code available at https://github.com/Lafayette-EshbaughSilveyra-Group/calibration-first

We introduce a calibration-first framework that produces supervision scores without access to ground-truth labels or a shared annotation space. Our framework aligns subset-specific scorers using a synthetic ordinal reference space before fusion. This reference space is constructed from ordered calibration features that represent the latent concept, providing a common scale on which otherwise incomparable scorer outputs can be aligned. Because our calibration procedure uses the reference space rather than training samples, it is independent of the training set's empirical distribution. Across three benchmark datasets, our framework consistently outperforms uncalibrated averaging and achieves higher primary-metric point estimates on the evaluation metrics than the best individual scorer. Performance relative to sample-dependent baselines varies by domain, with absolute differences below 0.02 on Ames Housing and below 0.01 on Breast Cancer Wisconsin and Wine Quality. After Bonferroni correction, differences remain significant for all three comparisons on Ames Housing and one on Breast Cancer Wisconsin. Additionally, we show that using fewer calibration levels per feature can closely approximate higher-resolution results at substantially lower computational cost. Together, these results support our framework as a viable approach to construct supervision scores when neither ground-truth labels nor a shared annotation space is available.

0.34

CircuitATLAS: Agentic reasoning over a systems neuroscience knowledge graph for target discovery in circuitopathies

cs.AI q-bio.QMq-bio.NC ·2026-10-07 ·Gabriel Ocana-Santero, Marko Tvrdic
Abstract · 22 pages, 5 figures, 4 tables

Drug discovery for neurological disease has traditionally centered on the molecules altered by disease. But the molecules that cause pathology are not necessarily the best points from which to reverse it. Here, we ask which otherwise unaltered molecular control points can be engaged to restore pathological neural circuits toward functional states. We present CircuitATLAS, a provenance-grounded systems-neuroscience knowledge graph and agentic framework for target discovery in circuitopathies. It structures literature-derived relationships across diseases, phenotypes, electrophysiology, circuits, brain regions, cell types and molecular effectors, while deliberately excluding direct disease-gene and disease-protein edges to reduce shortcut reasoning. The graph contains 3.83 million nodes and 7.66 million edges, including 5.31 million LLM-extracted relations, and incorporates structured datasets such as the Human Cell Atlas and new multimodal in vivo measurements. We then introduce an agentic workflow that reasons from measurable disease phenotypes through their circuit and cellular substrates to molecular interventions, therapeutic feasibility and clinical constraints. Finally, we introduce a human-governed in vivo lab-in-the-loop linking hypothesis generation to experimental iteration. Within this framework an agent nominated ATP1A3, the neuronal alpha3 Na+/K+-ATPase, as a control point on cortical excitability; interneuron-restricted expression of ATP1A3 abolished the beta- and gamma-band response to a focal 4-aminopyridine challenge in vivo, and the validated target was then carried into a structure-guided small-molecule campaign terminating in a defined assay to resolve the direction of modulation. CircuitATLAS thus provides a framework for discovering therapeutics based not only on what is molecularly disrupted in disease, but on what can be controlled to restore circuit function.

0.34

Layerwise Error Attribution for Fast and Robust Mixed-Precision Post-Training Quantization

cs.LG ·2026-10-07 ·Samy Houache, Yann Traonmilin, Jean-François Aujol
hot direction
Abstract

Mixed-precision post-training quantization is a network compression method that assigns bits layer by layer, under a global memory budget using a small calibration set. The main difficulties are to overcome the combinatorial nature of the allocation problem and to manage the sensitivity to small, potentially corrupted databases. Hence, an efficient allocation method should be fast to compute and preserve model quality when calibration data are corrupted. To design such a method, we derive a layerwise probabilistic analysis of the quantization error that separates propagated error from the local perturbation introduced at a given layer. We use this local term to build a separable score for a simple allocation algorithm, that requires no external solver. The probabilistic nature of our approach brings robustness to corrupted data. On denoising tasks with DRUNet, with an average budget of 4 bits per weight, our method matches or improves state-of-the-art mixed-precision baselines under clean calibration, and is more robust to corrupted calibration, with PSNR gains of up to 7.5 dB under the tested corruptions. Experiments show bit-allocation speed-ups from 28x to 2,570x over the studied baselines. For quantized diffusion models, our experiments show that a direct application of our framework also improves the state-of-the-art.

0.34

Sharp Asymptotic Theory of Maximum Likelihood Estimation for Gaussian Processes with an RBF Kernel

cs.LG math.STmath.PR ·2026-10-07 ·Ameer Qaqish, Sudipto Banerjee, Didong Li
hot direction
Abstract

Gaussian processes (GPs) are widely used across machine learning, spatial statistics, time-series analysis, optimization, Bayesian statistics, and scientific applications. A central component of a GP model is its kernel, which is typically specified through a parametric family. Among the most widely used choices is the radial basis function (RBF), also known as the squared exponential or Gaussian kernel, owing to its simple form, smoothness, and flexibility. In practice, the kernel parameters are routinely estimated by the maximum likelihood estimators (MLEs), as implemented by standard GP software. Despite this widespread use, the asymptotic behavior of the MLEs remains poorly understood under fixed-domain asymptotics, even for the RBF kernel. The main difficulty arises from the increasingly strong dependence among densely sampled observations and the nonlinear dependence of the covariance matrix on the kernel parameters. In this paper, we address this gap by providing, to the best of our knowledge, the first complete asymptotic characterization of the joint MLE of the spatial variance, lengthscale, and nugget variance under fixed-domain asymptotics. We establish consistency, derive convergence rates for all three parameters, prove joint asymptotic normality, and show that these rates are minimax optimal.

0.33

Nullify: Null-Space Activation Steering for Training-Free LLM Unlearning

cs.LG cs.AIcs.CR ·2026-10-07 ·Wei Zhai, Xiang Liu, Qiang Huang, Rui Qian, Lemao Liu +4 more
top venue
Abstract

Large Language Models (LLMs) inevitably internalize substantial amounts of sensitive or private information during pre-training, while LLM unlearning aims to selectively erase specific knowledge to prevent privacy leakage with minimal loss of model utility. However, existing methods struggle to balance forget quality with utility, and typically incur substantial computational costs due to parameter fine-tuning. To address this, we propose Nullify, a training-free, non-destructive activation steering method for LLM unlearning. Nullify employs steering vectors during inference to redirect privacy-related activations away from their memorized answers, while satisfying a null-space constraint that leaves retained-query activations essentially unaffected to maintain utility. Evaluations on TOFU and MUSE show that Nullify matches or surpasses established baselines in forget quality while achieving near-lossless preservation of model utility. By avoiding weight updates entirely, Nullify serves as an efficient, plug-and-play inference-time intervention framework.

0.33

WxFM-XL: Adapting Univariate Foundation Models to Multi-Station Weather Forecasting

cs.LG ·2026-10-07 ·Xiao Wang, Changjian Chen, Zhuo Tang, Rongwen Li, Hongwu Liu +1 more
hot direction
Abstract

With the rise of univariate time series foundation models (e.g., Sundial, Timer), initial efforts have been made to extend them to multivariate settings. However, these models mainly focus on modeling correlations among variables. When they are applied to multi-station weather forecasting, two important factors are often overlooked: (1) the spatial information of stations, and (2) different error priors of different stations relative to the foundation model. In this paper, we propose WxFM-XL, a model for adapting univariate time series foundation models to multi-station weather forecasting. WxFM-XL introduces a cross-station error correlation prior graph to capture stationwise error priors with respect to the foundation model. Building on this, we further propose a dynamic fusion mechanism that adaptively integrates a spatial correlation graph with the error correlation prior graph. Experiments on multiple datasets demonstrate that our model outperforms state of the art baselines.

0.33

For Those Who Believe in Faithfulness: Optimizing the Area Under Insertion and Deletion Curves for Ranking Relative Feature Importance

cs.LG cs.CV ·2026-10-07 ·Bjørn Leth Møller, Bulat Ibragimov, Christian Igel
code/data
Abstract

The adoption of machine learning for socially relevant tasks requires effective explainable artificial intelligence (XAI) methods to better understand the behavior of machine learning models. Attribution methods are a popular XAI approach in which input-output relationships are characterized by heat maps that reflect the relative importance of input features for a particular prediction. The quality of such maps is often assessed by measuring faithfulness based on the area under insertion and deletion curves, which measures changes in the model output as features are added and removed. In this study, we derive an objective function from this notion of faithfulness and a way to approximate its gradient. We establish the connection between insertion curves and top-$k$ feature selection, which leads to a loss function measuring the quality of attributions. Randomization of the loss allows us to efficiently approximate its gradient. To show the effectiveness of the general approach, we combine the loss function with the neural explanation mask framework. The resulting method, termed Ra-NEM, can be used with any differentiable model without affecting the model's performance. Experiments demonstrate that Ra-NEM provides accurate attributions robustly and efficiently. Compared to other algorithms, the attributions have not only higher faithfulness but also perform well in terms of other XAI metrics. The high inference speed of Ra-NEM makes the method suitable for online applications. The code is available online: https://github.com/baerminator/Ra_Nem

0.33

HAN-Mamba: Hierarchical Selective State Space Networks for Multi-Scale Financial Volatility Forecasting

cs.LG cs.CE ·2026-10-07 ·Mihai Bogdan Deaconu, Ioan Daniel Pop
top venuebenchmark
Abstract · 16 pages. Accepted for publication in Springer Lecture Notes in Artificial Intelligence (ICAART 2026 Revised Selected Papers). Extended vers…

Short-horizon realized volatility forecasting requires the integration of market information that evolves at incompatible temporal resolutions, from second-level order book dynamics to weekly regime drift. Our conference work introduced HAN-T, a hierarchical architecture in which scale-specific Transformer encoders process short, mid, and long-horizon streams and a learned attention fuser weighs their contributions. This article replaces the quadratic attention encoders with selective state space (Mamba) encoders while retaining attention only in the fuser, where the input is a three-token set rather than a long sequence. The resulting hybrid, HAN-Mamba, summarizes each stream through a recurrent state whose input-dependent gating matches two structural properties of volatility: persistent but decaying memory and abrupt regime shifts. On the Optiver Realized Volatility Prediction benchmark under time-aware five-fold cross-validation, HAN-Mamba improves mean RMSPE over HAN-T (0.1942 vs. 0.1965) with 33% fewer parameters. Its linear-time encoders further allow the high-frequency context to be extended from 60 to 240 buckets, reducing error to 0.1927 where the attention variant saturates, and support constant-time streaming updates at inference. Ablations attribute the gains to the encoder swap, confirm that the hierarchical prior transfers across sequence-model families, and show that the permutation-invariant attention fuser remains the correct mechanism for cross-scale integration.

0.33

MUNITE: Unified Multimodal Latent Inference for Any-to-Any Multimodal Generation

cs.LG ·2026-10-07 ·Kyeongmin Yeo, Minhyuk Sung
code/data
Abstract · Project page: https://munite-proj.github.io

We introduce MUNITE, a latent-variable framework for flexible any-to-any multimodal generation that treats encoding and latent generation as the same inference problem under different amounts of observed evidence. Given any subset of modalities, MUNITE models the conditional distribution over the latent representation associated with the complete observation. Full observation recovers deterministic encoding, no observation recovers the latent marginal, and intermediate subsets define conditional latent inference, all within a single conditional flow model. A shared latent sample captures variation that must remain consistent across generated targets, while modality-specific generative decoders model the remaining uncertainty independently. To learn these conditional distributions from incomplete training examples, we extend conditional flow matching through self-distillation: predictions conditioned on richer available observations supervise the same model conditioned on smaller subsets at the same intermediate latent state. When the richer-evidence trajectory follows the exact conditional flow, this provides the same expected learning signal as full-target denoising. Across PolyMNIST-D-Q, FFHQ64, and image-text-audio, MUNITE achieves competitive or better generation quality and source-target alignment, with higher joint-generation coherence. In particular, it attains the highest coherence in all one-to-many and unconditional image-text-audio comparisons, showing the effectiveness of unified latent inference across diverse multimodal settings.

0.33

AutoAdapt: Automatic Domain Discovery Enables Low-Cost Extensibility

cs.LG ·2026-10-07 ·Josh McGiff, Salma Mekaoui, Robert Shanahan, Nikola S. Nikolov
Abstract

Instruction-tuned models are deployed into environments where domains are heterogeneous and evolve, yet adding new domains or data typically requires costly retraining. We present AutoAdapt, a modular framework that incorporates new domains and data via targeted single-adapter training without modifying other adapters. The framework automatically discovers latent domains, uses them to train per-domain Low-Rank Adaptation (LoRA) adapters independently in parallel and performs parameter-free routing. Across 14 domain-specific benchmarks and GPT-4o pairwise judgements, AutoAdapt achieves parity with a LoRA adapter trained on all domains without requiring full-model retraining. We also find evidence of specialisation effect convergence across independent discovery methods. Overall, training each adapter on its own domain prevents domain interference by construction, thus enabling modular, taxonomy-free domain specialisation without aggregate performance loss or full model retraining.

0.33

Which Rollout Taught It That? BehaviorTrace and the Limits of Training-Data Attribution in Online RL

cs.LG cs.CL ·2026-10-07 ·Amit Nautiyal
code/data
Abstract · 11 pages, 2 figures, 4 tables. Code and data: https://github.com/AmitoVrito/BehaviorTrace

When reinforcement learning teaches a language model a new behavior, can we find the training rollouts that taught it? And when an attribution method says it can, how do we know the answer is real? We study both questions on online RL fine-tuning with GRPO, using a planted behavior with a known cause. We release BehaviorTrace, an open evaluation harness that combines full-gradient sketching, the planted-behavior setup, and controls for gradient magnitude, fluency, headroom, and variation across seeds and generation draws. Across three seeds on Qwen2.5-1.5B, much of the apparent attribution signal comes from confounds. A control that ranks training steps by gradient size alone, with no behavior target, reaches 4.2 to 4.5 times chance and matches or beats the best targeted estimator on two of three seeds. At saturated checkpoints, model fluency predicts the behavior label at least as well as every gradient method we compared it with. Once fluency is controlled, the per-rollout results change from seed to seed and from one generation draw to the next, so a single run cannot settle the question. One signal does hold on all three seeds. The gradient of the trigger tokens aligns with a target built where the behavior actually occurs. We turn these findings into a checklist for evaluating attribution in RL. We test existing estimators, including GAS (renormalized TracInCP) and a TRAK-style estimator, and do not propose a new one.

0.33

Conformal Prediction under Partial Verification

cs.LG stat.ML ·2026-10-07 ·Zijun Yu, Yu Gu, Vahid Partovi Nia, Masoud Asgharian
Abstract

Conformal prediction provides prediction sets with finite-sample guarantees, but the label verification required for calibration can be expensive. We develop a partial verification method that returns exactly the same prediction sets as complete verification. We characterize calibration certificates, the verified information sufficient to determine the conformal threshold, and design a procedure that coordinates verification across calibration examples. For finite thresholds at high coverage, its verification cost is less than twice the minimum certificate cost when candidates are checked in order. Across retrieval, mathematical solutions, and configuration evaluation, it reduces verification cost by 15-82% compared with verifying calibration examples one at a time, while producing identical prediction sets.

0.33

Multimodal LLMs Can Learn to Read Brain Signals: A Vision--Language Model for Unified Multi-Task EEG Decoding

cs.LG cs.AIcs.CV ·2026-10-07 ·Parastoo Azizeddin, Omid Sharafi, Maryam M. Shanechi
Abstract

Learning EEG representations that generalize across cognitive tasks, subjects, and recording conditions remains a key challenge in electroencephalography (EEG) decoding. Recent advances in foundation models have improved EEG decoding performance, yet a fundamental open question remains: how to effectively interface neural signals with these models to enable multi-task learning across datasets. To investigate this question, we introduce BraVista, a visual-language framework that encodes multichannel EEG signals as structured images and enables multi-task learning through instruction-conditioned vision-language models (VLMs). Our approach relies on continued post-training of a general-domain VLM, leveraging its visual and linguistic priors to adapt to neural signals without a separate large-scale EEG-specific pretraining stage. We evaluate BraVista on four datasets spanning sleep staging, emotion recognition, cognitive workload classification, and abnormal EEG detection, showing strong performance across these tasks. Further analyses show that the choice of EEG-to-image representation is critical to performance. Moreover, through controlled perturbations of the EEG signal, we observe a gradual performance degradation under increasing noise, suggesting that the model relies on EEG-relevant information rather than superficial visual patterns. Together, these findings establish structured visual representations as an effective and scalable interface between neural signals and general-domain foundation models for unified multi-task EEG decoding.

0.33

EvoSignal: LLM-Guided Evolutionary Design of Modular Traffic Signal Control Programs

cs.LG ·2026-10-07 ·Leizhen Wang, Peibo Duan, Zhenlin Qin, Yancheng Ling, Jian Xu +3 more
code/data
Abstract

Effective traffic signal control (TSC) requires policies that respond to changing traffic demand and network conditions while meeting different control objectives. However, adapting existing strategies often involves repeated manual design and adjustment, making it difficult to systematically explore better control rules for a target network. Large language models (LLMs) can automate this process, but directly using them to select signal phases leaves decision rules embedded in black-box models and incurs recurring inference costs and latency. This paper formulates TSC as a modular program design problem and proposes EvoSignal, an LLM-guided evolutionary framework using traffic knowledge and performance feedback. The modular representation separates traffic feature extraction, local phase prioritization, and optional network-based priority adjustment. Starting from several established strategies, EvoSignal improves programs through feedback on congestion and signal operation, retaining strategies with different performance trade-offs. The resulting programs operate without online LLM inference. Simulation experiments across five scenarios on two real-world road networks show that the selected default EvoSignal program reduces waiting time by 16.8--49.2\% relative to the lowest waiting time achieved by the 20 conventional, reinforcement learning-based, and LLM-based baselines in each scenario. A program prioritizing travel time and queue length outperforms all 20 baselines on all three metrics in the search scenario and remains among the top three on each metric when transferred unchanged to the other four scenarios. These findings support automated design of inspectable control programs that transfer across the evaluated road networks and traffic demands. Code is available at https://github.com/georgewanglz2019/EvoSignal.

0.32

A Society of Researchers: Designing Institutions for Populations of Autonomous Research Agents

cs.MA cs.AI ·2026-10-07 ·Ali Asaria, Deep Gandhi, Tony Salomone
positionhot direction
Abstract · 4 pages. Short version of A Society of Researchers: Institutional Design for Populations of Autonomous Scientific Agents, doi:10.5281/zenodo…

Deployments of research agents are moving to populations of thousands that share one pool of compute, while most current systems organize one project at a time or leave the population unorganized. We argue that such a population will acquire an organization whether or not its designers provide one, so designers should provide it explicitly, and that the multi-agent systems community holds the tools to do so. We propose a society of agents, a population of persistent agents under explicit institutions, and develop it for science as a society of researchers built on six principles. Principal investigators compete for compute through requests for proposals, independent review, and grants; a human governor, the mayor, allocates resources and assigns no tasks. In a running society of ten thousand researchers, asked only to improve the pretraining of language models, one lab reported a way to reach the same quality with about 30% less compute, a result the labs that tested it do not yet agree on. We close with six open problems for the agents community.

0.32

Decoupling Exploration from Optimization in RLVR

cs.LG cs.AIcs.CL ·2026-10-07 ·Saif Punjwani, Micah Goldblum
code/data
Abstract · 20 pages, 16 figures, 9 tables. Code: https://github.com/SaifPunjwani/Exploration-Distillation. Checkpoints: https://huggingface.co/SaifPunj…

Modern language models undergo reinforcement learning with verifiable rewards (RLVR) on top of already-trained checkpoints. A key promise of RLVR is the discovery of new reasoning strategies. In principle, a model can sample novel ideas absent from its prior training data. In practice, however, augmenting RLVR with strong novelty incentives has seen limited success and can degrade model quality. Because verifiable rewards supervise only a narrow slice of the model's knowledge and behavior, such degradations are difficult to recover from. Instead, we decouple exploration from optimization in a framework we call Exploration-Distillation (ExpDis). We train one or more explorer policies with a novelty bonus in the reward, filter their trajectories for correctness and quality, and distill them into a separate student policy. The student policy is then trained without a novelty bonus. We repeat the above procedure for several rounds, alternating between exploration and optimization. This decoupling allows us to aggressively scale exploration without degrading the student policy. Across seven mathematical reasoning benchmarks and two model families, ExpDis outperforms DAPO at the same wall-clock budget. Moreover, we observe improved pass@$k$ scaling, indicating that ExpDis produces models that generate more diverse correct solutions.

0.32

Low-rank tensor structure of precipitation and its application to satellite-reference merging

cs.LG physics.ao-ph ·2026-10-07 ·Ryan Solgi, Rohan Shankar, Hugo A. Loaiciga
Abstract

The intermittent and variable nature of precipitation makes its accurate estimation over extended domains difficult, yet its spatiotemporal structure suggests that a low-rank representation may be possible. This work represents daily precipitation over the contiguous United States (CONUS) as spatiotemporal tensors and applies CANDECOMP/PARAFAC factorization, showing that preserving the native spatial and temporal modes yields more accurate reconstruction than factorizing independent daily fields or unfolded space--time matrices. Building on this finding, this work presents TMerge, a tensor-based framework that integrates satellite precipitation with sparse reference observations through shared low-rank spatial and temporal factors. TMerge was applied to correct the IMERG Final Run product with climate prediction center reference observations over CONUS. During 2019-2022, TMerge increased correlation from 0.53 to 0.85 and reduced root-mean-square error and mean absolute error by 48.2% and 29.3%, respectively. TMerge consistently outperformed linear bias correction, quantile mapping, and neural networks across seasons, precipitation-intensity regimes, and regions. Improvements were spatially coherent and largest in coastal regions where IMERG errors were greatest. These results demonstrate that low-rank tensor structure parsimoniously approximates the dominant spatiotemporal variability of precipitation and provides a practical mechanism for improving satellite estimates under limited reference observations over extended domains.

0.32

Visible Reasoning Is Not a Universal Optimizer: Persona- and Thinking-Dependent Effects in Analytics Code Generation

cs.LG cs.AIcs.SE ·2026-10-07 ·Bhawani Shankar Leelar, Pawan Chorasiya, Davin Hill, Robert E. Tillman, Tamer Soliman
benchmark
Abstract · 26 pages, 8 figures, 16 tables

Visible Chain-of-Thought (CoT) is often treated as a broadly useful reasoning instruction, yet analytics code generation combines natural-language ambiguity, schema grounding, target-language constraints, and model-specific inference behavior. Because the same analytics request can be expressed in two distinct target languages-SQL and Python (pandas)-this setting provides a natural test of a common but under-examined assumption: that visible reasoning is more effective when its representation matches the requested target, as in "think in SQL" or "think in Python." Together with generic instructions such as "think step-by-step," such recommendations remain insufficiently evaluated under controlled, execution-based comparisons. We study a query matched SQL-pandas benchmark that crosses persona phrasing, target language, visible-CoT format, control prefixes, direct generation, and internal-reasoning configurations. The results do not support either a universal accuracy advantage from visible CoT or a consistent benefit from matching the reasoning representation to the target language. Instead, the effects depend on the persona, target, model configuration, and internal-reasoning setting. The control ablations further distinguish effects of reasoning content from those of prompt format. These findings indicate that reasoning strategies should be selected jointly for the model, persona, target, and internal-reasoning configuration rather than adopted as universal defaults. More broadly, the study provides a controlled framework for identifying when visible reasoning improves executable generation, when it primarily perturbs model behavior, and when the internal-reasoning configuration is the more consequential factor.

0.32

Estimating Uncoded Crash Factors with Tabular Foundation and System One Models: Kumo Tabular and Jev

cs.LG stat.APstat.ME ·2026-10-07 ·Amir Rafe, Subasish Das
code/data
Abstract · 26 pages, 9 figures, 8 tables. Code: https://github.com/pozapas/kumo-jev-crash-records

Road safety programs count the coded fields of police crash records, while the officer's narrative, which often records factors the fields omit, is rarely read. A safety office thus cannot tell how much its counts miss or where to review. This study develops and evaluates a system that joins both views of the 5,601,890 Texas crashes from 2017 to 2025 into population estimates with stated validity. An in-context tabular foundation model, Kumo Tabular, reads the coded record of every crash, a calibrated System One model, Jev, reads the narratives of two probability samples, and human judgments recalibrate its probabilities. A multiwave predict-then-debias estimator joins the three tiers, and a second human tier drawn with recorded probabilities checks the estimates by design. For hydroplaning, medical episodes, fatigue, animals, and phone use, the narrative documents more injury crashes than the coded field, 15,074 against 7,340 for phone use, and the human check agrees with all fifteen estimates within its margin. A re-read list ranked by Kumo Tabular finds confirmed discordance 7 to 58 times as often as random reading. At the planning cost of human coding, one further round of human judgments would cut the root mean square relative half-width from 22.0 to 16.2 percent, against 21.2 for reading every narrative. Two calibrated readers of different views, joined by a sampling design, give a safety office counts, a discordance map, a validated re-read list, and a reading budget, with Kumo Tabular reading the table at 15 times the speed of TabPFN 3.5.

0.32

Learning How to Search for Plans with Exponentially Less Space

cs.AI ·2026-10-07 ·Dominik Drexler, Simon Ståhlberg, Markus Fritzsche, Blai Bonet
Abstract

Heuristic search for a plan can store exponentially many states, even when its heuristic is almost perfect. We instead learn search control, one specification per domain, written as an indexical policy: a generalized policy with registers that hold objects and modes that sequence its rules. We add the choose rule, which loads an object into a register and marks a backtracking point, where one candidate suffices; every other rule must work for all of its outcomes and needs no search. Our main result is that structural termination, which rules out infinite executions, also bounds every execution by a polynomial in the number of objects. A depth-first procedure then finds a plan in polynomial space, however large the state space, with no list of visited states. The cost is time, exponential only in the choice depth, the number of real choices along an execution. Any class that such a policy solves therefore lies in NP, and in P at constant choice depth. We learn these policies with a language model in a counterexample-guided loop that certifies termination, verifies the training tasks, and keeps the choice depth small. With the learned policies, the procedure solves 1,709 of 1,890 test tasks of the IPC 2023 Learning Track and the Autoscale Agile suite, more than LAMA, BFWS, and Levitron, and most of them within one second and 100 MiB.

0.32

Learning infinite context windows in recurrent architectures via spatial neural computing

cs.LG eess.SYq-bio.NC ·2026-10-07 ·Aleix Salvador-Pomarol, Arthur N. Montanari, Earl K. Miller, Adilson E. Motter, Jorge Cortés
Abstract · 12 pages, 4 figures

Recurrent neural networks (RNNs) offer linear-time scaling with sequence length while requiring only constant memory, yet they struggle to capture long-range dependencies due to vanishing gradients and limited receptive fields. To address these limitations, we introduce a second-order recurrent model in which the standard neuron-to-neuron communication is replaced by a spatially evolving field governed by (discretized) partial differential equations. Drawing inspiration from the role of cortical waves in brain computation, this mechanism allows structured spatiotemporal patterns to serve as an implicit, high-capacity memory. We show that the resulting model is equivalent to a structured infinite-order RNN in which the current state depends explicitly on its entire history of past states, yielding an effectively unbounded receptive field with a fixed number of parameters. We further derive constructive conditions to ensure marginal stability, constraining the gradient spectrum on the unit circle and thereby eliminating vanishing and exploding gradients. Empirically, the proposed architecture outperforms other recurrent models on long-horizon benchmarks while using substantially fewer parameters, demonstrating that spatial dynamics can effectively bridge the gap between efficient inference and long-term memory.

0.31

MemoWM: How World Models Change What Agents Need to Remember

cs.LG cs.AI ·2026-10-07 ·Bingfan Zeng, Zhisheng Chen, Chenbo Sang, Zhengwei Xie, Jinpeng Wang +4 more
code/data
Abstract

Long-term agents face growing storage demands as they accumulate experience. World models capture reusable regularities that can reduce the information stored for each experience. We formulate the problem of memory allocation conditioned on a world model and introduce MemoWM, a framework that uses shared predictions to compress retained information and reconstruct omitted content. Its task-aware allocation rule balances the expected impact of reconstruction errors against storage cost, retaining information with downstream value beyond the predictive prior. Across five long-term agent-memory benchmarks, MemoWM achieves 42.42\% average answer accuracy, exceeding the strongest baseline by 2.62 percentage points, while reducing average experience-specific storage by 53.9\% relative to MIRIX, the most storage-efficient baseline. Further analysis shows that stronger world models reduce per-experience storage at comparable task quality. Accounting for model parameters reveals a trade-off between shared model capacity and recurring storage costs, with the capacity that minimizes total storage increasing as more interactions are retained. Our code is available at https://github.com/Feld-maxiu/MemoWM.

0.31

SPD-MetaFormer is what you need for small-data brain decoding

cs.LG stat.ML ·2026-10-07 ·Zhida Wang, Wei Lyu, Guo Yu, Sui Tang
Abstract

Brain signal decoding is challenging because neural recordings are noisy and vary across individuals, while labeled data are often limited. Recent attention-based models on the symmetric positive definite (SPD) manifold have nevertheless achieved strong performance using covariance and connectivity representations, yet the contribution of learned token weighting remains unclear. We examine two representative architectures, MAtt (based on log-Euclidean geometry) and GBWAtt (based on generalized Bures--Wasserstein geometry), and find that their learned attention weights remain close to uniform after training. We relate this behavior to bounded similarity parameterizations that, under the original softmax scaling, limit attention-weight contrast. Moreover, replacing learned weights with uniform weights, throughout training and evaluation, has little effect on mean predictive performance while preserving each model's original aggregation geometry. Motivated by these findings, we introduce SPD-MetaFormer, an attention-free architecture built on uniformly weighted Fréchet aggregation under log-Euclidean geometry. Its backbone uses a geodesic residual to update a summary token and a shared spectral feedforward map to transform all tokens, followed by a learned weighted readout. Token states remain SPD-valued until tangent-space classification. Across three EEG benchmarks, SPD-MetaFormer achieves competitive results relative to published Euclidean and manifold baselines. Separate matched reproductions test learned versus uniform weighting within MAtt and GBWAtt. These results suggest that, in the short-sequence and limited-data regimes studied, carefully designed SPD architectures can provide a simpler and effective alternative to adaptive manifold attention.

0.31

Sequential Pretraining Favors Large Models

cs.LG ·2026-10-07 ·Mohnish Harwani, Yujia Zheng
Abstract

Large neural networks often acquire capabilities that small models fail to learn. Does this stem from large models learning more representative features, or from being more robust to unaccounted-for adverse effects introduced during training? We define and quantify one such adverse effect, primacy bias, as the extent to which exposure to early data distributions impairs later learning. We show that small models can allocate learning capacity inefficiently toward early distributions, whereas sufficiently overparameterized models are robust to this effect. This inefficiency is particularly consequential in pretraining, where foundation models often encounter heterogeneous data distributions sequentially rather than jointly. As a result, small foundation models can struggle to learn distributions encountered late in training, which is particularly harmful when later data emphasizes desirable capabilities such as code, mathematics, and reasoning. Motivated by these findings, we introduce Exposure Therapy (ET), a simple regularization that promotes more efficient allocation of learning capacity during sequential pretraining. We demonstrate that ET improves foundation models' performance on late data distributions as well as overall capability in models up to the billion-parameter scale. Overall, our results suggest that some benefits of large foundation models may arise from greater robustness to adverse training effects, rather than from learning more representative features, and that improved training algorithms can recover some of these advantages in smaller models.

0.31

Self-attention summary networks for subsurface velocity-model building from common-image gathers

cs.LG physics.geo-ph ·2026-10-07 ·Shiqin Zeng, Yunlin Zeng, Abhinav Prakash Gahlot, Zijun Deng, Felix J. Herrmann
Abstract

Common-image gathers (CIGs) contain physically meaningful information about velocity-model errors through reflector focusing and residual moveout, but in conventional imaging workflows they are typically used only as diagnostic tools. In this work, we propose a multiscale self-attention summary network that maps high-dimensional 3D CIG volumes into compact conditioning embeddings for probabilistic subsurface velocity inversion. These learned embeddings preserve offset-dependent kinematic structure and spatial coherence while reducing variability caused by background-velocity mismatch. Conditioned on these summary embeddings, a flow-matching model learns a transport from a Gaussian source distribution to the posterior distribution of plausible velocity fields. Numerical experiments show that, compared with direct conditioning on raw CIGs, the proposed summary network improves posterior velocity inference. In particular, the multiscale attention design provides greater robustness to background-model mismatch, yielding more accurate posterior reconstructions and lower predictive uncertainty.

0.31

Evaluating Rubric Generation with Interventional Transfer

cs.LG ·2026-10-07 ·Erik Skalnes, Layne C. Price, Raviteja Anantha, Michael Oberst
positioncode/data
Abstract · 29 pages, 6 figures. Code and cached model outputs: https://github.com/oberst-lab/evaluating-rubrics-with-interventional-transfer

Instance-specific rubrics are common in AI benchmarks where reliable evaluation requires specific expert knowledge. This approach is difficult to scale, prompting research into the generation of rubrics with large language models (LLMs). However, even when expert rubrics are available as references, it is unclear how to productively evaluate the quality of generated rubrics at scale. In this paper, we introduce a method for the evaluation of rubric generation, which we call Interventional Transfer (IT), based on the idea that two rubrics are similar if they move together when a response is perturbed to pass/fail one of them. In contrast to existing approaches for evaluation of rubric generation, we argue that different forms of interventional transfer can be used to evaluate the utility of generated rubrics for different tasks. For instance, we apply this approach in a case study on HealthBench, where we demonstrate an asymmetry in rubrics generated by Qwen3.8-27B, Deepseek-V4-Flash, and Opus-5, used to evaluate responses from GPT-5.6-Terra. Perturbations that degrade responses according to the generated/expert rubric transfer into lower scores on the corresponding expert/generated rubric, but perturbations that improve on one rubric do not reliably transfer into higher scores on the other. We argue that this finding has implications for the usage of LLM-generated rubrics for performance monitoring and hill-climbing. We contrast our approach with existing approaches for rubric evaluation, which do not surface the same asymmetry that we observe.

0.31

Speaking the Navigator's Language: Trajectory-Grounded Instruction Translation for Frozen Aerial VLN Agents

cs.AI ·2026-10-07 ·Xi Chen, Zhe Liu, Xiaogang Xu, Jiafei Xu, Chunyi Zhou +6 more
Abstract

Aerial vision-and-language navigation (VLN) agents are typically trained on detail-rich, trajectory-aligned commands, whereas users issue short, intent-driven instructions; on a frozen OpenFly navigator, this \emph{instruction gap} drops success rate (SR) from $31.03\%$ to $11.33\%$. To scale translator training, we prompt a language model with human-written style examples to convert original commands into paired, intent-centered Weak commands, which yield $15.27\%$ SR. We introduce the \textbf{Trajectory-Grounded Instruction Translator (TGIT)}, a front-end that keeps the navigator frozen and translates Weak inputs into agent-executable commands by learning from its trajectory outcomes. The resulting Weak-trained translator raises Weak-input SR to $37.93\%$ and transfers zero-shot to real human instructions ($11.33\%{\rightarrow}32.51\%$); it also improves held-out OpenFly ($4.95\%{\rightarrow}20.79\%$) and yields recovery on CityNav and AirVLN.

0.31

Adaptive Multi-Discriminator WGAN Framework for Resource-Constrained Internet of Vehicles Using Reinforcement Learning and Game Theory

cs.LG cs.DC ·2026-10-07 ·Farhoud Jafari Kaleibar, Amr M. Zaki, Marin Litoiu
Abstract · 15 pages

Managing machine learning workloads as a network service introduces a resource-orchestration problem distinct from conventional model training; which nodes should be allocated to a task, how communication and computation budgets should be divided among them, and how service quality should be sustained as connectivity and node availability change with mobility. Deploying Generative Adversarial Networks (GANs) in Internet of Vehicles (IoV) environments is a demanding instance of this problem; resource constraints, dynamic network topologies, and competing optimization objectives mean that traditional GAN architectures cannot simultaneously achieve high accuracy, efficient resource use, low delay, and low communication overhead. This paper introduces an adaptive multi-discriminator Wasserstein GAN (MD-WGAN) framework that integrates reinforcement learning with game-theoretic coordination to address these challenges jointly. In our framework, roadside units host generators paired with Deep Q-Network (DQN) agents that select discriminator subsets and manage distributed training across mobile vehicular nodes, while a game-theoretic coordination step allocates training epochs between generators and discriminators. A unified optimization objective ties adversarial learning quality to resource efficiency, communication overhead, and latency under vehicular constraints, allowing the framework to continuously adapt its training behavior as network conditions change. Evaluation on real-world NGSIM trajectory data shows that the framework attains prediction accuracy comparable to state-of-the-art GAN baselines - the lowest RMSE (1.029) and MAE (0.894) among all evaluated methods - while markedly improving resource efficiency: average CPU utilization is reduced by roughly 28% and mean memory usage by roughly 6%, at competitive communication overhead and latency.

0.30

ProtocolMatch: Protocol-Dependent Model Selection for Scientific Dynamics Forecasting

cs.LG quant-ph ·2026-10-07 ·Lu Wei, Yufeng Wang, Haibin Ling
Abstract · 14 pages, 4 figures, 8 tables

Scientific dynamics forecasting is often framed as an architecture choice, although deployment is also determined by observed history, rollout feedback, compute budget, physical objective, and test distribution. We formulate protocol-dependent model selection and introduce ProtocolMatch, a compute-matched, validation-selected, and failure-preserving evaluation framework. On driven quantum-spin dynamics, we compare recurrent, patched-attention, causal-attention, and low-rank linear predictors across three independently generated datasets. The causal-attention--recurrence ordering reverses as the training set grows within a fixed two-spin task, while a linear predictor has the lowest mean error in the six-spin local-observable comparison. Restricting observed history worsens every refreshed-history view but improves every closed-loop view in the four-spin study. A latest-state MLP has lower error than persistence on every dataset under state refresh across all five cells, yet its closed-loop rank varies by system and includes finite explosive errors. Physical penalties improve targeted consistency without reliably improving prediction error, and in-distribution intervals lose most coverage after a driving-frequency shift. Thus scientific model selection should return a predictor with its protocol and report accuracy, physical validity, and shifted-distribution reliability separately.

0.30

KDFP: A first-principles approach to knowledge distillation in large language models

cs.LG ·2026-10-07 ·Ryan Swift, Konstantinos Psounis
code/data
Abstract · 30 pages, 6 figures; Code available at https://github.com/ryanaswift7/kd_first_principles

Knowledge distillation is an established technique for improving the capabilities of small, efficient student models by training them with the representations of larger, more capable teacher models. Much of the recent work in the distillation of large language models (LLMs) has focused on distilling abilities learned during post-training, such as instruction following, chain-of-thought reasoning, and tool usage. This has left a large research gap in general knowledge distillation for LLMs, which is essential for developing efficient and private systems suitable for deployment on edge devices. We take a first-principles approach, evaluating previous lessons from prior works and conducting new explorations to develop a distillation methodology suitable for modern LLMs. We present KDFP, a novel methodology for white-box general knowledge distillation in LLMs. We demonstrate that KDFP outperforms existing methods by 1.6% $-$ 4.9% across 9 benchmarks while increasing training efficiency by up to 99.1% through ephemeral parameter reduction.

0.30

When Should an In-Context Learner Expand Its Hypothesis Space?

cs.LG ·2026-10-07 ·Weihan Li, Xinlei Chen, Junhao Wu, Tianshi Zheng
Abstract · 46 pages, 12 figures, 13 tables

Learning systems adapt quickly within a familiar family of models. The harder step is deciding, from a prediction failure, whether opening a richer family is worth its cost. We cast this as a costly sequential decision and introduce the Structural Revision Environment (SRE), which produces matched failures from noise, exceptions, change within the family or structure outside it, varies the price of expansion and the remaining horizon independently of the evidence, and admits an exact normative solution of the one-shot decision. The solution shows that revision is a value boundary, not an evidence threshold: except at the lowest prices, the richer family becomes the more probable explanation long before expanding becomes optimal. Small Transformers trained in SRE reproduce the expansion boundary from utility alone. Pretrained 7B and 8B language models fail at a definite point: their predictions carry a failure-sensitive signal that their decisions do not reflect. Told the gain of expanding, they weigh it against price and horizon only when they reason, and the one model that finishes its reasoning on the history alone still does not estimate that gain. Controlled post-training of the small Transformers moves the prior and the sharpness of predictions, and neither moves the criterion.

0.30

GAGR-Lab: Evaluating Joint Spatial-Geometric and Analytic Function Reasoning

cs.AI cs.CLcs.CVcs.LG ·2026-10-07 ·Jingyao Zhang, Yun Li, Lu Han
Abstract · 15 pages, 1 figure, 7 tables

Joint spatial-geometric and analytic function reasoning requires translating a perceived spatial configuration into a symbolic function whose executed curve satisfies geometric constraints. We present GAGR-Lab, a framework for measuring this capability through Cartesian game scenes, explicit function semantics, and authoritative Rust trajectory execution. It distinguishes spatial perception, metric grounding, geometric relations, function interpretation, function construction, and constrained synthesis. We specify four configurable scene-difficulty presets and a prospective 24-cell diagnostic design, while reporting only the subset actually evaluated. A bounded pilot of one hosted model (Llama 3.2 11B Vision Instruct) using two API credentials as execution replicas yields 72 balanced games with 432 attempts, 429 valid provider responses, and no target hits; exploratory ordinary-function prompt variants also fail to hit, while the structured localization interface yields no scoreable outputs. A privileged analytic search control independently succeeds on 600 directional cases from 300 generated scenes, with exact repeatability and 1,200 successful vertical-reflection or translation checks. The framework separates serving reliability, symbolic compliance, and geometric success, and preserves exact model-visible inputs and realized paths. A staged protocol outlines diagnostic calibration, held-out replication, multi-model comparison, and paired robustness tests. The contribution is an operational research framework with an executed pilot and a clearly identified prospective study plan; the full difficulty matrix and comparative model results remain untested.

0.30

Self-Organization from Constrained Geometric Radiation

cs.LG ·2026-10-07 ·Ming Lei
theoryunusual
Abstract · 13 pages, 4 figures

How does dynamic order emerge spontaneously in closed systems without external driving? Existing paradigms all require external energy flows, temperature quenching, or slow driving. Here we report constraint-induced self-organization via geometric radiation in coupled metric evolution systems. Simulations reveal a universal four-stage cycle: stress accumulation, super-exponential radiation, chaotic collapse, and convergence to a fractal limit cycle, a novel attractor topology we term the wedge-shaped attractor, with five quantized curvature states and fractal micro-fluctuations. We identify four jointly sufficient conditions: an irreversible geometric horizon, persistent stress injection from quantum coherence, endogenous geometric tension between incompatible curvatures, and effective fluctuations. Their synergy triggers a critical avalanche at the horizon boundary. We prove three theorems: the Geometric Horizon Theorem, the Geometric Energy Dissipation Theorem (implying wave-like entropy evolution in closed systems), and the Radiation as Phase Transition Channel Theorem. We further establish the Constraint-Induced Self-Organization Theorem: these conditions guarantee the complete cycle with probability one. Systematic scans reveal a critical noise threshold and power-law scaling of radiation onset. We verify universality across 12 configurations, multiple noise types, and three geometric flows. This work establishes a new paradigm for closed-system self-organization, forging an exact mathematical duality between classical nonlinear constraints and gravitational horizons.

0.30

Computations of the slice genus and the unknotting number of links via machine learning

cs.LG math.GTstat.ML ·2026-10-07 ·Yutong Dai, Oliver Hayman, András Juhász, Ludovico Morellato
unusual
Abstract · 72 pages, 34 figures

Links are disjoint unions of circles smoothly embedded in $S^3$. We use reinforcement learning and Bayesian optimisation to obtain new upper bounds on several link invariants that are not known to be algorithmically computable: the slice genus and the unknotting number for links, and the strong slice genus for algebraically split links. We also compute lower bounds using known invariants. Combining the upper and lower bounds, we obtain new exact values in many cases. Our unknotting agents can reproduce the non-additivity of the unknotting number for several counterexamples due to Brittenham and Hermiller, in some cases finding new unknotting trajectories.

0.30

Beyond the Ergodic Wall: A Discrete Geometric Physics Sandbox for Analysing AI Scaling Limits and Complexity Collapse

cs.LG cs.AI ·2026-10-07 ·Simon Richard Daniel
unusual
Abstract · 17 pages, 8 core pages including tables, 9 pages references tables (updated references to bbl)

This paper exposes the ergodic ceiling and thermodynamic inefficiency of current deep learning, which converges to a statistical average of historic human knowledge. True semantic novelty requires a path-dependent, spatiotemporally bounded observer (a Data LifeCone) to inject non-ergodic insight, achieving KL divergence and avoiding manifold lock-in. AI Safety must recognise that a mature Artificial Superintelligence (ASI) would regard human-AI symbiosis as a thermodynamic necessity to avoid model collapse. We therefore propose hard physical containment via a digital physics sandbox powered by a Holographic E8 Projection Engine to verify models against real-world constraints. Spacetime is modeled as an information substrate of nested face-centered cubic (FCC) lattices of oscillating Planck-scale spheres maximizing local information and entropy density. Cut-and-project methods from the E8 root lattice produce a quasi-crystalline geometry where tetrahedral voids support SU chiral structure and elastic-shear eigenvalues generate candidate mass spectra. Rest mass is treated as discrete, integer microstate counts on local holographic boundaries (Bekenstein bound), replacing floating-point approximations with strict integer arithmetic to provide an information-theoretic definition of matter. Stable particles emerge as recurring lattice dislocations, and continuum recovery proceeds via variational renormalisation-group flows and Fourier Neural Operators that learn continuous spectral operators to recover the Schrödinger equation as an emergent statistical description. Crucially, these top-down topological constraints offer a mechanism for "NP-to-P" complexity collapse: by restricting an algorithm's proposal space to physically conserved causal trajectories, the sandbox prunes the combinatorial tree to deterministic, polynomial-time paths.

0.30

Transformed Samplers with Variance Reduction

cs.LG stat.MLstat.ME ·2026-10-07 ·Siran Liu, Michalis Tisias, Petros Dellaportas
Abstract

Markov chain Monte Carlo (MCMC) methods are the standard tool for computing expectations under complex probability distributions. Control variates reduce the variance of the resulting estimates, but a good control variate requires solving the Poisson equation of the sampler, which rarely admits a closed-form solution. Exact solutions are available when the sampler's kernel has a known spectral decomposition on a simple reference density. In our work, we extend these solutions to general targets through a learned change of variables. A bijection, such as a normalizing flow, is trained so that the target becomes close to the reference in a latent space, and we show that Markov kernels and their Poisson solutions are transformed by any bijection. Running such samplers in the latent space then yields explicit control variates, and the estimator is consistent under mild tail conditions on the map and target. Importance sampling (IS) from the flow is the limiting case of the same construction and the control variates apply to it as well. Experiments on synthetic targets and real posteriors compare the procedure against state-of-the-art samplers and control variates.

0.29

ASPIRE: Saddle-Point Discovery through Set Prediction and Physical Refinement

cs.LG ·2026-10-07 ·Yucheng Zhao, Quanyou Zhang, Shaoxiang Qin, Haixuan Xu, Xiongye Xiao
benchmark
Abstract

Predicting thermally activated diffusion and defect evolution with event-driven models requires identifying atomic rearrangement mechanisms and their activation barriers. Discovering the associated saddle points is a major computational bottleneck: multiple rearrangements may originate from one metastable state, while costly local searches can fail or repeatedly converge to the same saddle. To address this challenge, we introduce ASPIRE (Atomistic Saddle-Point Inference with Refinement for Events), a framework that predicts a set of saddle candidates from a single initial atomic environment and refines them through Dimer searches on the original interatomic potential. The framework's equivariant set predictor, Ev-Quiformer, integrates (i) geometry-conditioned scalar-vector event slots for generating multiple saddle-point proposals and (ii) a decoder that maps each slot to a full atomic displacement field by combining atom, slot, and anchor-relative vectors with invariant coefficients. We also contribute two datasets: (i) BCCFE4VACAV-4000, comprising 4,000 four-vacancy body-centered cubic iron configurations and 65,450 reference events grouped by initial state for set supervision and post-refinement evaluation; and (ii) BCCFE-1TO4VAC, comprising 5,372 configurations with one to four vacancies each. Theoretically, we establish conditions for proposal equivariance. Experimentally, ASPIRE achieves 77.20% reference-event coverage on this benchmark, compared with 75.73% for a conventional Dimer baseline, while requiring approximately half as many Dimer force evaluations. In a timing evaluation on 50 configurations, ASPIRE reduces wall time per configuration from 478.8 s to 176.3 s under the stated hardware settings.

0.29

Activation-Aware Weight Tensorization: A Calibration-Time Preconditioner for Tensor-Network LLM Compression

cs.LG cs.AI ·2026-10-07 ·Alessandro Beatini, Marco Maronese, Emanuele Rodolà
Abstract

Post-training tensor-network compression replaces Transformer linear layers with Tensor Train (TT) or Tree Tensor Network (TTN) operators, but standard decompositions minimize weight-space Frobenius error rather than functional error under the layer's activation distribution. We propose Activation-aware Weight Tensorization (AWT), a training-free calibration wrapper that preconditions each weight matrix with a diagonal activation-derived scale before an unchanged TT/TTN solver and deploys the result with only an input-side elementwise rescaling. Across Llama 3.1 8B, Ministral 8B, and Qwen2.5 7B, AWT consistently improves vanilla TT/TTN tensorization at 2-6 times compression: under single-operator replacement, AWT closes 12-35% of the WikiText perplexity gap to the dense baseline across the three model families and 2-6 times compression settings; while under multi-operator Llama suffix replacement it closes 27-60% across attention-group and all-seven-matrix settings. The gains also transfer to downstream HellaSwag and ARC-Challenge evaluations. We further show that diagonal preconditioning is a robustness-modularity tradeoff rather than a diagonal-covariance assumption: a dense full-covariance oracle wins its own weighted objective in 80/81 cases, yet diagonal AWT gives better held-out functional fidelity in 53/81 cases. Together, these results position AWT as a principled, modular preconditioner for improving functional fidelity in fixed TT/TTN compression pipelines without modifying the decomposition solver.

0.29

Cognitive Schemas, Laws and Tasks

cs.AI ·2026-10-07 ·Antal Jakovác, András Telcs
Abstract · 25 pages , 1 table

This paper asks how explicit representations can support reusable cognitive schemas in knowledge-based problem solving. We develop a structural framework in which schemas are organized by the information and relations required for their use, rather than introduced as unrelated primitives. The framework also distinguishes context-dependent relations from more stable structures that can be reused across different representations. Tasks are described through the information available, the unknowns to be determined, and the constraints that admissible solutions must satisfy. This makes it possible to separate limitations of the representation from limitations of the solving procedure. In particular, we distinguish inconsistency, underdetermination, and contextual insufficiency, where the current representation lacks distinctions or relations required by the external task meaning. We also show formally when a reduction of representation preserves the task-relevant solution structure. The resulting task--schema interface offers a structured way to describe representational conditions relevant to problem solving. It supports the reuse and stabilization of derived knowledge while remaining independent of the particular mechanism used to generate candidate solutions. This may provide a useful component for future solver architectures that combine structured knowledge, verification, and learned proposal mechanisms.

0.29

Stream-Based Active Learning with Cooperative Neural Networks for Data-Efficient Partial Inverse Design: An Automotive Glass Run Channel Case Study

cs.LG cs.AI ·2026-10-07 ·Agung Nugraha, Hyerin Kwon, Heungjun Im, Gian Antariksa, Jihwan Lee
Abstract · 21 pages, 12 figures, 3 tables

Inverse design in engineering often runs into a simple problem. Each labeled training sample must be produced through expensive simulation, so building a large dataset is slow and costly. This study addresses that problem for partial inverse design, where only some design variables are specified and the rest must be inferred to reach a target performance value. We propose CoNN-AL, a framework for data-efficient partial inverse design that adds stream-based active learning to the Cooperative Neural Network with Denoising Autoencoder (CoNN-DAE). The model estimates predictive uncertainty through Monte Carlo dropout and uses it to decide, in real time, which incoming candidate samples are worth labeling, so the limited labeling budget is spent on the most informative designs. We validate the framework on a real-world automotive glass run channel dataset of more than 900,000 unique simulated designs. With only 20,000 actively selected labels, about 2.3% of the training pool, CoNN-AL reaches R-squared values of 0.967 to 0.982 across all missing-variable levels, approaching the upper-bound models trained on far more data. It reaches R-squared of at least 0.95 with 30 to 40% fewer labels than random sampling at the more difficult missing-variable levels and, at the most challenging level, is the only strategy in this study to reach R-squared of 0.98. Together with this work, we publicly release the dataset to support future research on data-driven design.

0.29

Residual Learning in Empirical Asset Pricing

cs.LG q-fin.ST ·2026-10-07 ·Dexin Peng, Xiaoyu Wang
Abstract · 58 pages, 8 figures

Shallow models are special cases of deep models, and deep models theoretically have the potential to outperform the shallow ones. However, the existing empirical asset pricing literature provides strong benchmarks for shallow models. Residual learning allows neural network models in asset pricing to go deeper by preserving and refining their shallow counterparts. The out-of-sample Sharpe ratio for value-weighted long-short portfolios of deep residual models (2.07) is higher than that for the corresponding shallow ones (1.92) and more than twice that of the deep feedforward models (0.89). We show that model depth is a source of additional economic value in asset pricing. Residual learning can be used to deepen other neural-network-based asset pricing models if they contain intermediate layers. Our design also provides one way to scale asset pricing models, making native "large asset pricing models" more feasible.

0.29

Derivative Gaussian Processes on a Two-Direction Budget

cs.LG stat.ML ·2026-10-07 ·Hyunseok Seung, Matthias Katzfuss
Abstract

Gradient observations promise more accurate Gaussian process (GP) surrogates, but the cost of incorporating them has long stood in the way of realizing that promise. We propose a derivative GP with a budget of just two directions per observed gradient. One direction focuses on each gradient's direct contribution to target prediction, while the other aggregates its indirect contributions through correlations with the conditioning function values. Within a Vecchia approximation, where each prediction conditions on $m$ nearby inputs in $d$ dimensions, this construction represents their $md$ gradient coordinates using at most $2m$ directional derivatives, giving $\mathcal{O}(m^3)$ dense factorization cost per prediction target. For general conditioning sets, we bound the posterior approximation error relative to using full gradients and characterize when the error is small or the approximation is exact. In simulations, our method matches the accuracy of a leading exact gradient-reduction method at equal conditioning set size. Because its cost grows much more slowly with that size, it can use conditioning sets well beyond the memory limit of the exact method, reaching lower prediction error with a small fraction of the time and memory. Notably, our method can exploit gradient observations while requiring less computation time or memory than function-only GP baselines.

0.29

Sparsifying Stochasticity, Not Capacity: Partial Stochasticity via Deep Weight Factorization of Prior Scales

cs.LG stat.ML ·2026-10-07 ·Marius P Linhard, Maurizio Filippone
Abstract

Bayesian neural networks need not be fully stochastic to be universal conditional density approximators, but it remains open which parameters should be stochastic. We learn this split by applying deep weight factorization to the prior scales, which are the standard deviations of the parameter priors, while fitting the functional prior to a Gaussian process with a maximum mean discrepancy objective. A parameter whose prior scale falls below a cutoff becomes deterministic and is optimized during inference, so the regularizer sparsifies stochasticity rather than capacity. We give a certificate for universal conditional density approximation that is checkable in linear time, together with a minimal repair when it fails. We further show that the common hybrid scheme of sampling some parameters and optimizing the others is stochastic approximation for a type-II maximum a posteriori objective, and that coupled step sizes can leave a tracking error that does not vanish as the step size shrinks. On a bimodal target, the learned split stays close to an unconstrained reference across all budgets and is insensitive to the cutoff, while random masks that distribute the same prior scales across layers are worse by up to two orders of magnitude. On UCI benchmarks, our method performs on par with a fully stochastic network while keeping about half of its parameters deterministic.

0.29

What Does the Effective Dimension of a Hybrid Quantum Neural Network Actually Measure?

cs.AI quant-ph ·2026-10-07 ·Md Masum Mia, Amith Khandakar, Mohamed Arselene Ayari
Abstract · 21 pages, 13 figures, 11 tables; includes 3 pages of supplementary material

Hybrid quantum neural networks (QNNs) pass features from a classical backbone into a variational quantum circuit, and many studies log a scalar called the effective dimension as a sign of capacity or model quality. The effective dimension is usually defined as the participation ratio (PR) of the quantum Fisher information matrix (QFIM), but studies in practice log a PR of cheap parameter sensitivities instead, and the meaning of that scalar is rarely tested. We show that the logged scalar equals the PR of Pauli-Z output sensitivities, that each sensitivity is at most n times its QFIM diagonal entry on n qubits, and that a PR of the QFIM diagonal overstates the full-matrix PR by an exact factor set by the off-diagonal mass. We then test the scalar on CIFAR-10 and EuroSAT with four frozen backbones, one of them a random projection, over 4 to 16 qubits, 1 to 4 layers, three entangling patterns, and three seeds. At a fixed circuit, CIFAR-10 accuracy spans 23.14% to 86.44% while the scalar stays between 6.63 and 9.13, and its correlation with accuracy changes sign across seeds in 7 of the 14 configurations. Depth raises the share of active parameters from 0.39 to 0.91, near a ceiling we derive, and initial gradient variance falls by a factor of 34.5 from 4 to 16 qubits. Removing entanglement raises accuracy by 5.3 to 8.2 points at the same parameter count. The effective dimension therefore describes circuit geometry rather than model quality, so capacity, trainability, and accuracy each need their own diagnostic.

0.29

Closed-Form Noise Calibration Against Membership Inference for Random-Allocation DP-SGD

cs.LG cs.CR ·2026-10-07 ·Murat Bilgehan Ertan, Marten van Dijk
Abstract

DP-SGD protects training data by adding Gaussian noise to clipped gradients. The amount of noise is usually chosen by running a numerical privacy accountant inside a search. We study DP-SGD with random allocation, where each epoch uses every record once, at a randomly chosen step. For this setting we give a one-line formula that bounds the accuracy of every membership inference attack (MIA) on the trained model. With $M$ steps per epoch, $E$ epochs and noise multiplier $σ$, and with membership and non-membership equally likely a priori, the attack accuracy is at most $\frac12+\frac14\sqrt{(1+(e^{1/σ^2}-1)/M)^E-1}$. The formula comes from the chi-square divergence between a Gaussian distribution and a Gaussian mixture that dominates random allocation. It is interpretable and gives $σ$ in about a microsecond. Where applicable, our formula needs at most about half the noise of the state-of-the-art closed-form bound. To measure how close the bound is, we also derive an exact expression for the attack accuracy of these two distributions and evaluate it numerically. Calibrating to this exact expression requires $13.0\%$ to $20.2\%$ less noise than the formula in our main experiments, and since it is exact, no accountant that knows only $M$, $E$ and $σ$ can certify a smaller $σ$. In training, the resulting $σ$ outperforms the formula and matches a published accountant in test accuracy. It is found in seconds and certified in minutes, whereas every search we ran with that accountant took longer or returned at least $0.62\%$ more noise. We show that MIAs on the trained models stay below the bound.

0.29

Bias-Aware Adjoint Calibration and Optimal Control of Stochastic Multiscale Bioprocess Digital Twins

cs.LG stat.ML ·2026-10-07 ·Keilung Choy, Wei Xie
Abstract · 35 pages, 8 figures

We develop a bias-aware digital-twin calibration and control framework for multiscale bioprocess models within a biological systems-of-systems (Bio-SoS) paradigm. The digital twin is represented by a stochastic differential equation (SDE) model and calibrated from sparse observations using quasi-likelihood estimation and adjoint sensitivity analysis. SDE generator-based moment expansions characterize truncation-induced parameter bias, while forward-backward adjoints quantify how calibration uncertainty propagates to value functions and policy performance. The resulting parameter-error distribution supports both policy-directed data acquisition and uncertainty-aware policy optimization through a second-order Gaussian-averaged objective. We characterize the asymptotic behavior of the resulting exploration criterion and derive a physical-system performance under the optimized policy. To implement these ideas, we develop an Actor-Simulator algorithm that jointly updates model parameters, selects informative experiments, and optimizes control policies. Numerical studies demonstrate improved calibration accuracy, sample efficiency, and control performance relative to state-of-the-art baselines.

0.29

Validity Without Ground Truth: What Stated-Preference Economics Offers the Evaluation of Language Models

cs.AI cs.CLecon.GN ·2026-10-07 ·Daniel Robert Kling Alexander, Catherine Louise Kling
position
Abstract

Many of the questions now put to large language models have no correct answer to score against: what a policy is worth, which option a user should choose, how to weigh competing values. Stated-preference economics has faced this problem for decades. It judges survey responses without knowing the true value, through a framework of validity and related concepts: content, construct, and criterion validity, reliability, incentive compatibility, and consequentiality. We argue that this framework is a general method for evaluating language models, and we set out what each concept means for LLM evaluation. We demonstrate the approach using a published water-quality stated preference economic valuation survey (Vossler et al. 2023) administered to six models. In this economic application, the validity tests take the form of predictions from economic theory: demand should slope down, and willingness to pay should respond to the scope of the good and to income. The tests separate the models sharply. Two older models fail the most basic test at a household income level of \$75,000, and the two newest pass every test of theoretical validity we can score, but diverge on convergent validity. Passing validity tests shows that a model's answers are coherent, not that they are correct.

0.29

Know the Shape, Find the Fault: Topology-Conditioned Diagnosis of Multi-Agent LLM Failures

cs.MA cs.AI ·2026-10-07 ·Xinwen Liu, Zhuocheng Pan, Isabella Zhu, Jawei Zhang, Xudong Liu +1 more
Abstract

Multi-agent LLM systems coordinate task execution through exchanges of information among agents. When coordination breaks down, similar symptoms in execution traces can reflect different problems in how information is passed, used, or verified. Communication topology captures how agents exchange information and provides structural cues for distinguishing coordination failure modes. Using these cues for diagnosis requires establishing how topology relates to failure patterns and recovering the relevant structure from execution traces that lack explicit topology labels. We analyze the relationship between communication topology and failure patterns and introduce MAScope, a two-stage framework for topology-conditioned diagnosis. Its Trace Structural Extractor TSE recovers communication topology from heterogeneous execution traces by grounding an interaction graph in message evidence. The Topology-Conditioned Judge TC-Judge then classifies failures using the trace, predicted topology, an empirical failure prior estimated from separate labeled traces, and a short description of topology-specific failure patterns. Under a fixed orchestration structure, the recovered topology can be reused across executions. Experimental results show a statistically significant association between communication topology and failure type, with $χ^2 = 409.9$ and $p = 1.2 \times 10^{-70}$. On the \num{851} MAST-clean traces, ground-truth topology context raises gpt-mini's Macro-F1 from $0.173$ to $0.350$. With predicted topology, the pipeline achieves $0.346$, approaching the trace-only gpt-5.4 baseline of $0.372$. For \num{1000} traces under a fixed orchestration structure, the projected pipeline cost, including one topology extraction, is approximately $6\%$ of repeated gpt-5.4 diagnosis cost. These results show that topology-conditioned context improves failure diagnosis and supports lower-cost deployment.

0.29

Coefficient Calibration as Selection Pressure in Symbolic Regression

cs.LG ·2026-10-07 ·Mattia Billa, Veronica Guidetti, Federica Mandreoli
Abstract · 18 pages, 8 figures, 7 tables

In memetic symbolic regression, candidate structures are compared after coefficient calibration, so the calibration protocol itself contributes to evolutionary selection. Standard centralized calibration evaluates each structure at its pooled-sample optimum, ignoring how stable this calibration is under covariate shifts, and can thus favor structures whose fit relies on sample-specific coefficients. We propose Dirichlet-Sinkhorn Constant Averaging (DSCA), a calibration strategy that partitions the optimization data into equally sized subsets with different covariate distributions, calibrates each candidate independently on every partition, and evaluates it at the mean of the resulting parameters. We show that the excess loss of DSCA relative to centralized calibration vanishes at the population level for correctly specified, identifiable expressions, whereas under misspecification it persists when partition-specific calibrations do not aggregate to the pooled optimum. On synthetic benchmarks and ten real-world datasets, DSCA improves functional recovery and the accuracy-complexity trade-off over centralized Broyden-Fletcher-Goldfarb-Shanno and Levenberg-Marquardt calibration, under selection by negative log-likelihood and by the Akaike and Bayesian information criteria. Mechanism analyses associate the DSCA excess loss with the generalization gap and show that the effect is not reproduced by repeated centralized fitting. These results indicate that controlled heterogeneous calibration provides a complementary source of selection pressure in symbolic-regression search.

0.28

From Log-Odds to Shapley Values: An Explanatory Geometry for the Weighted Naive Bayes Classifier

cs.AI stat.MLcs.LG ·2026-10-07 ·Vincent Lemaire, Fabrice Clérot
Abstract · 15 pages

This paper studies the construction of an explanatory space for a weighted naive Bayes classifier from the supervised representation induced by the model. We start from the classical supervised distance based on conditional log-likelihoods and introduce a discriminative reformulation based on log-odds, which is more directly related to the classification decision. We then show that this representation induces a distance that exactly coincides with the $\ell_1$ distance between vectors of analytical Shapley values, thereby providing a formal explanatory interpretation of the geometry induced by the model. Finally, we empirically compare several supervised distances derived from these representations using a $k$-nearest neighbors classifier. This work highlights a close link between supervised distance, local explanation, and predictive behavior, from a primarily methodological perspective.

0.28

Extreme Binary Classification: Extreme Value Theory for Extreme Constraint on False Negative

cs.LG stat.MLstat.APstat.ME ·2026-10-07 ·Samuel Gruffaz, Muhammad Fawad, Jaakko Nevalainen
Abstract

While binary classification is one of the most extensively studied problems in machine learning, the regime in which the goal is to learn a classifier with an almost zero false negative rate remains largely unexplored. In this paper, we introduce the Extreme Binary Classification problem, where the objective is to learn a classifier whose false negative rate $α$ is constrained by $ε_{N_1}=o_{N_1\to\infty}(1/N_1)$, with $N_1$ denoting the number of positive examples in the training set. To address this problem, we propose a threshold adaptation method theoretically grounded in guarantees derived from Extreme Value Theory, together with a feature selection procedure based on a permutation test applied to sample maxima. Experimental results on four real-world datasets of varying sizes demonstrate that our approach compares favorably with state-of-the-art methods. In addition, we illustrate its interpretability through an application to a cancer screening dataset.

0.28

Shared Low-rank Basis Factorization for Data-free Mixture-of-Experts Compression

cs.LG cs.AI ·2026-10-07 ·Tianxiao Cao, Jiahe Shao, Yuning Qiu, Kyohei Atarashi, Hisashi Kashima +1 more
top venue
Abstract · Accepted to Findings of EMNLP 2026

Mixture-of-Experts (MoE) large language models decouple capacity from compute through sparse routing, but their large parameter count creates storage and serving challenges. We analyze three MoE compression families: expert pruning, expert merging, and weight reconstruction, and derive structural error bounds showing that pruning and merging can incur non-vanishing errors tied to routing and expert heterogeneity. In contrast, weight reconstruction avoids these structural costs by preserving expert structure and routing. Motivated by the analysis, we propose Shared Low-rank Basis Factorization (SLBF), a data-free weight reconstruction method that uses rank-$k$ bases shared among experts, enabling richer cross-expert sharing, faster convergence, and lower reconstruction error. A post-hoc gauge fixing removes redundant parameters at no representational cost. Across five MoE architectures spanning 16B to 122B parameters, SLBF consistently outperforms methods from all three compression families.

0.28

Few-Shot Learning for Personalised Automated Pain Assessment

cs.LG cs.AI ·2026-10-07 ·Heinke Hihn, Ibrahim Eisawy, Patrick Thiam, Hans A. Kestler, Friedhelm Schwenker
benchmark
Abstract

Pain perception varies substantially across individuals, making it difficult for population-based classifiers to generalise across all subjects in a dataset. One way to account for subject variability is to train personalised classifiers. In this work, we evaluate Few-Shot Learning, a sub-area of Meta-Learning, as an approach to personalisation in automated pain assessment. We re-interpret the shift from population-level to subject-level evaluation as a task-domain shift, where the observed classes remain fixed but the target subject changes. We evaluate our method on the BioVid Pain Database, the SenseEmotion Database, and the PainMonit Experimental Dataset (PMED), reaching 85.75% and 35.49% accuracy on BioVid and 82.37% and 41.88% on SenseEmotion in the binary and multi-class settings under a Leave-One-Subject-Out CV protocol respectively, and 90.47% on PMED, for which only a binary benchmark exists. Using samples to implement k-shot conditioning, the accuracies can be improved to 86.25%, 40.06%, 83.43%, 44.08%, and 91.25%, respectively. To further evaluate the effects and robustness of our method, we provide additional ablation experiments and investigate the personalisation effects. Our results suggest that support-conditioned few-shot adaptation can improve average performance under inter-subject variability.

0.28

Shared Gaussianization: What Gaussian Regularizers Certify About Contrastive Learning, and What They Miss

cs.LG stat.ML ·2026-10-07 ·Ruoyu Zhao, Yuting Chen, Jinheng Zhang, Zhehao Zou, Tong Che
Abstract · 27 pages, 4 figures, 3 tables. Ruoyu Zhao and Yuting Chen contributed equally; Tong Che is the project lead

What can a distribution-matching regularizer such as SIGReg in LeJEPA certify about contrastive learning? We study shared Gaussianization (SG), a characteristic-function Gaussianity test on the average of two normalized views, scaled by an independent $χ_d$ radius. Because disagreeing views shorten the average, one test detects both misalignment and non-uniformity. SG vanishes exactly at the aligned, uniform minimizers of population InfoNCE, and under equal marginals it bounds the InfoNCE excess by $4\cdot 3^{3/4}β$ times the square root of the SG loss, plus a term linear in the loss. The square-root rate and this dimension-free constant are sharp, and no squared mean-embedding distance on view pairs achieves a faster rate. With an explicit alignment term, a rotation-invariant uniformity test gives a linear bound if and only if its spectrum dominates that of InfoNCE's kernel $e^{βu^\top v}$; SG's own test does, Gaussian kernels $e^{-γ\|u-v\|^2}$ qualify exactly when $γ\ge β/2$, and moment matching never does. Away from the optimum, the objectives differ. Along an isotropic nuisance channel, pure SG lowers its loss by adding per-view nuisance whenever the shared code is non-uniform. An alignment weight above the channel's gain makes the nuisance-free solution a strict local minimizer; for LeJEPA, the same rule gives a critical SIGReg weight that decreases with the batch size. At finite batch size, an off-diagonal U-statistic removes a plug-in bias toward misalignment. In controlled latent-variable models, pure SG retains per-view style, an alignment weight above the measured gain removes it, and for LeJEPA at three batch sizes the measured gain separates the encoders that retain style from those that do not. InfoNCE training also reaches a lower SG$_{0.2}$ loss than SG$_{0.2}$ training from scratch, which points to an optimization gap.

0.28

RunningTab: Direct Workspace Interaction with Environment-Side Tabs

cs.AI cs.CL ·2026-10-07 ·Jinheon Baek, Soyeong Jeong, Yumin Choi, Dongsu Han, Sung Ju Hwang
Abstract

Much knowledge work produces new deliverables from files a workspace already holds, and LLM agents are beginning to take such work over. Through direct corpus interaction, an agent can search and read any of those files from a terminal with no indexing, and producing a deliverable from many of them in this way is what we call direct workspace interaction (DWI). Reaching the files, however, is only half the task: nothing keeps track of what the task asks for, what has been read, and what was listed but never opened, all of which slip through the context window without leaving a trace, so an agent may extract a figure and still deliver a report without it. To address this, we present RunningTab, a framework that equips direct workspace interaction with an environment-side tab: a per-task record of what the task still owes, kept by the environment alongside the agent. Specifically, the agent adds its requirements, while the environment records every file read as an excerpt with its provenance and every listed but unopened file as a candidate; the agent can then see each requirement beside its best-matching excerpts and top unopened candidates, resolve it against matching content or set it aside with a reason, and, should it try to finish with requirements still open, receive them in a finish check. We validate RunningTab on three benchmarks with three LLMs, where it consistently outperforms plain DWI and baselines that keep the record in the model, while its tab usually holds the values a deliverable needs once seen.

0.28

A Tale of Two Error Categories: Exploring Concealed Trade-Offs in the Errors of Automated Judges in Evaluation of Uncertainty Quantifiers

cs.AI ·2026-10-07 ·Evgenia Ilia, Wilker Aziz
published
Abstract · Accepted at INLG 2026

The wide adoption of LLMs across broad NLG applications heightens the importance of providing users with the means to avert errors and hallucinations. Uncertainty quantification is poised to fill that gap; with low uncertainty (high confidence), as a proxy for correctness, allowing users to be selective (e.g., reject low-confidence, likely incorrect responses). Correlation between confidence and correctness then serves as a useful criterion for evaluation of uncertainty quantifiers (UQs). But in NLG, where diverse responses can be adequate to a prompt, obtaining reliable correctness judgements is not simple, especially without human intervention. Errors in automated judgement are hardly avoidable and known to diminish the reliability of evaluation protocols (Santilli et al., 2025; Ielanskyi et al., 2025). In a meta-analysis of published work, we show that automated judgement is the present norm. Besides, automated judgements are rarely validated against human ones, and the validation of the UQ evaluation they automate is even rarer. With experiments in question answering, using 4 LLMs, human and automated judgements and 7 popular UQs, we find that i) a judge's performance can only coarsely predict the observed impact of its errors on the reliability of UQ evaluation, and that ii) judgement errors tend to misrepresent informative UQs most. We link these observations to patterns of correlation between confidence and categories of judgement error.

0.28

Policy Learning with Weak Signals

cs.LG stat.MEmath.ST ·2026-10-07 ·Benedikt Koch, Winston Chou, Aurélien Bibaut, Nathan Kallus
Abstract

Policy learning in digital experimentation faces three challenges: weak signal-to-noise ratios, rich covariate spaces, and massive data volumes. We formalize this regime by modeling treatment-effect estimates from increasingly fine covariate partitions as Gaussian observations with bounded signal-to-noise ratios. We establish that, in general, the optimal treatment policy is not learnable in this setting. Even learning the optimal policy value suffers from impractically slow rates. However, when treatment effects vary smoothly, we derive minimax-adaptive policies based on linear smoothers that achieve vanishing welfare regret. We demonstrate the practical value of our framework by applying it to large-scale real-world experiments at Netflix, showing that personalized linear-smoothing policies can dominate unpersonalized policies even in this challenging empirical setting.

0.28

Average-Reward Reinforcement Learning for Multichain MDPs: A Hierarchical Decomposition Approach

cs.LG math.OC ·2026-10-07 ·Huizhen Yu, Isaiah Heidt
Abstract · 60 pages, 4 figures

We study learning optimal policies in average-reward multichain Markov decision processes (MDPs), where the optimal gain may depend on the initial state and recurrence structures vary across policies, creating challenges for reinforcement learning (RL) methods. We propose an asynchronous value-iteration-based RL algorithm that requires no model knowledge beyond the MDP's transition graph and leverages Bather's decomposition to hierarchically partition the state space into communicating subsystems and transient states. This decomposition induces a recasting of the global decision problem into structured subproblems, which our algorithm exploits. We show that the algorithm converges to the optimal gain and produces gain-optimal policies after finite time. Building on this base algorithm, we develop two further algorithms: one approximately solves the multichain average optimality equations to obtain near gain-optimal policies, and another targets near bias-optimality by approximating the optimal bias function and solving an induced average-reward multichain MDP using the base algorithm. We provide almost-sure convergence guarantees for all three algorithms and empirically compare their tradeoffs, showing that the latter two also consistently improve transient performance relative to the base algorithm. To our knowledge, these are the first essentially model-free average-reward RL algorithms for general multichain MDPs without reductions to discounted problems.

0.28

When does a network's training history predict its future learning better than its current state? Evidence from a response probe and a forecasting screen

cs.LG ·2026-10-07 ·Martin Hofmann, Patrick Mäder
Abstract · 18 pages, 6 figures, 5 tables

Networks that behave alike now can still learn differently when training continues. Work on loss of plasticity and critical periods shows that the path to a state shapes what follows; it does not show whether the path carries information that a measurement of the state itself misses. We ask when the training history of a network predicts its future learning better than its current state. In a main study, small multilayer perceptrons were trained under three history regimes (42 histories), and future learning was measured at four checkpoints by a short probe: a copy of the network trained for 100 updates on a new task. Before the prediction result was read, the protocol checked the probe. It responded monotonically to a function-preserving rescaling of hidden units, repeated measurements agreed (intraclass correlation 0.940, [0.903, 0.997], in the least reliable class, mean of three repeats), and a re-initialisation of units was visible directly after it but not 100 to 200 updates later. A history state of at most four dimensions did not improve on a calibrated model of the current state (gain -21.4%, 90% interval [-91.9, 8.1]; required in advance: 10%). A companion screen on 1,560 synthetic regression runs asked the same question for a target further away, the final error of the run. There, history models forecast better than the current validation error after 12 of up to 240 epochs (compact state 30.3%, [15.8, 39.4], a contextual comparison) and were not distinguishable from it after 48. In both studies the history was informative only while the current state was not yet informative about the target; this reading was formed after the results.

0.27

TRACE: A Governance Framework for Measuring Explainability Debt in Production AI Systems

cs.LG cs.AI ·2026-10-07 ·Harish Kant Pathak
published
Abstract · 28 pages, 4 figures, 4 tables. Keynote presented at ICIDS 2026 (Manipal University Jaipur) and IEEE Al-Khwarizmi 2026. IEEE Senior Member #1…

Production AI systems deployed in high-stakes domains accumulate a governance liability that existing monitoring frameworks fail to detect: the progressive inability to explain individual decisions when regulators, auditors, or affected individuals demand accountability. We introduce TRACE (Transparency, Risk, Accountability, Compliance, and Explainability), a seven-instrument governance framework for measuring, tracking, and remediating Explainability Debt in production AI systems. The foundational instrument, the Explainability Debt Score (EDS), quantifies the proportion of production decisions falling below a governance-defined explainability confidence threshold at any point in time. Complementary instruments include DART (Debt Accumulation Rate Tracker for breach forecasting), SHIV (Scenario Health and Integrity Validator for daily governance), FDE (Feature Drift Evaluator for causal attribution), HVE (Human Validation Engine), AIDE (Audit Intervention Decision Engine), and ZERO (Zero Explainability Risk Optimiser for remediation). Through a twelve-month longitudinal case study of a production fraud detection system processing 50,000 daily financial transactions, achieving 98.46% accuracy and ROC-AUC of 0.9990, we demonstrate that an EDS of 0.23 on audit day was statistically predictable six months in advance using DART trajectory analysis (beta = 0.008/week, R-squared = 0.94, 95% CI: [0.006, 0.010]), and that 78% of Explainability Debt was concentrated in the highest-regulatory-risk decision category (transactions above $10,000), a risk asymmetry completely invisible to system-level metrics. TRACE provides the first quantitative operational architecture for EU AI Act Article 13 compliance in production AI deployment, establishing a new subdiscipline of explanation governance distinct from explanation generation.

0.27

NEMORA: Neural Equivariant Multipole Operators for Long-Range Atomistic Learning

cs.LG physics.chem-phphysics.comp-ph ·2026-10-07 ·Jay L. Kaplan, Samuel Varner, Rebecca Willett, Juan J. de Pablo
Abstract · 59 pages, 5 figures, including supplementary material

Equivariant graph neural networks have emerged as foundational architectures for machine-learned interatomic potentials, approaching quantum-chemical accuracy at a fraction of the computational cost. These models describe local atomic environments accurately, but finite spatial cutoffs truncate long-range information flow, and stacking message-passing layers can lead to over-smoothing and over-squashing. Existing long-range extensions either prescribe a fixed analytical propagation kernel, restrict long-range communication to scalars or degree-preserving channels, are only approximately equivariant, or incur super-linear computational cost. Combining learnable long-range equivariant transport with multiscale many-body expressivity and efficient scaling for larger systems remains a central challenge. We introduce Neural Equivariant Multipole Operators (NEMORA), a neural equivariant extension of the Fast Multipole Method (FMM) for learning long-range tensorial representations. NEMORA generalizes the FMM's analytical multipole expansion and translation operators to learned equivariant counterparts on an adaptive spatial hierarchy. Its operators couple angular degrees and form many-body interactions across length scales, retaining the FMM's hierarchical organization and analytical radial factors as physical inductive biases while learning data-dependent long-range couplings. NEMORA evaluates in linear time and memory complexity, allowing it to treat larger systems than other long-range methods reaching hundreds of thousands of atoms, and it augments both symmetry-constrained and unconstrained short-range backbones. On non-local benchmarks, it reduces force and energy errors relative to the short-range backbones by over an order of magnitude and up to three orders of magnitude, respectively, which is better than or competitive with existing long-range extensions in accuracy.

0.27

MotherTree: Meta-learning on synthetic data improves decision tree training

cs.LG ·2026-10-07 ·Ziyuan Wang, Fredrik D. Johansson
hot direction
Abstract

Conventional decision tree algorithms produce effective, transparent models that can be audited, communicated, and deployed independently of the training data, but require learning every new task from scratch. In contrast, tabular foundation models demonstrate that meta-learning from a synthetic prior distribution enables strong in-context prediction for previously unseen tasks, especially in small-sample regimes. However, this approach does not produce a standalone model that can be inspected in isolation. We introduce MotherTree, a tabular transformer that meta-learns decision tree induction: given a training set for a new task, it outputs a hard, axis-aligned decision tree, equivalent in form to classically trained trees, in a single forward pass. MotherTree is pre-trained on a synthetic prior using stochastic gradient descent without requiring reference trees for supervision. On established benchmarks with controlled sample size, the approach is competitive with size-matched trees from common algorithms: recursive partitioning, gradient-based tree learning, globally optimal trees, and distillation from tabular foundation models. Notably, MotherTree consistently improves over from-scratch gradient-based learning and acts as a strong initializer: task-specific tuning of the generated tree outperforms the corresponding from-scratch learner on all benchmarks and sample sizes. These results show that meta-learning can provide effective inductive biases for learning stand-alone, small decision tree classifiers.

0.27

What a Reporting Convention Hides: A Matched-Budget Audit of Quantum Natural Gradient with an Exactly Computed Metric

cs.LG quant-ph ·2026-10-07 ·Lu Wei, Yufeng Wang, Haibin Ling
Abstract · 29 pages, 6 figures, 12 tables. Lu Wei and Yufeng Wang contributed equally

Several published comparisons of variational quantum optimizers time only runs that reach a target loss, or read the verdict at a single target. Either convention could decide whether an optimizer's costlier steps pay off. We measure how much each convention changes verdicts among Adam, simultaneous perturbation stochastic approximation (SPSA) and quantum natural gradient (QNG), on initializations held out from the selection of settings. We compute the exact metric that preconditions QNG, price every step in circuit evaluations and give every method the same budget. In a median pooled over circuit widths, cost families and a sweep of settings with common misses, SPSA needs more than twice Adam's evaluations to reach a loose target. Dropping the censored runs that miss the target hides this gap. On the global-cost family we hold fixed the settings selected for a strict target. QNG then usually reaches the loose target after Adam but the strict target first. QNG's strict-target lead disappears when the metric's simulator price, linear in the number of parameters, is replaced by an assumed hardware count quadratic in that number. We recommend charging every miss the budget and reporting verdicts across targets.

0.27

Beyond Prediction: A Component-Level Study of Concept Recovery in In-Context Learning

cs.LG cs.AI ·2026-10-07 ·Xufeng Liu, Fang Wan, Yi Liu
Abstract

In-context learning (ICL) enables language models to make accurate predictions from demonstrations, but prediction success alone does not reveal the extent to which the underlying concept has been recovered. Existing evaluations typically assess recovery at the level of the whole concept, making it difficult to identify which components are recovered and to separate model limitations from ambiguity in the demonstrations. We introduce a component-level framework for concept recovery that decomposes concepts into their defining components and uses controlled structural interventions to evaluate both component-level and whole-concept recovery under identifiable demonstrations. Across four instruction-tuned LLMs and three composition forms, we find that models often achieve high prediction accuracy despite poor whole-concept recovery, with systematic over-specification of relevant attributes and level constraints. Individual structural components are generally insufficient for whole-concept recovery, whereas jointly specifying complementary components, particularly relevant attributes and their level constraints, substantially improves recovery. However, this improvement is fragile: a single plausible but unnecessary level constraint can substantially disrupt recovery. Finally, joint supervision of recovery and implicit prediction substantially improves recovery while preserving prediction performance, with the learned capability transferring across composition forms. Overall, our component-level analysis provides a more diagnostic view of what models recover from demonstrations than the existing holistic view.

0.27

Agentic AI-Assisted Modeling for Production Scheduling: Assessment in Constraint Programming

cs.AI cs.SE ·2026-10-07 ·Ángel Sánchez-Fernández, Javier Pernas-Álvarez, Diego Crespo-Pereira
hot direction
Abstract · 29 pages, 9 figures

Developing optimization models for production scheduling requires substantial expert effort. Research on large language models (LLMs) has followed two directions: specialized approaches for automated modeling, mostly for mixed-integer linear programming, which often rely on dedicated training or problem-specific architectures that limit industrial deployment; and agentic artificial intelligence for operational decision support, which generally assumes that the optimization model already exists. This study bridges both directions by assessing whether general-purpose LLMs, orchestrated as agents without task-specific training, can formulate and implement constraint programming models from natural-language problem descriptions. Singleagent and multi-agent architectures are integrated with a Model Context Protocol server that provides context-aware retrieval of solver documentation to mitigate hallucinations during implementation. Both are compared with a direct LLM baseline on six industry-oriented problems covering flow-shop, job-shop, flexible job-shop and resource-constrained warehouse scheduling, using three LLMs and assessing modeling accuracy, execution success, latency and token consumption. Formulation proves largely within reach of current LLMs, whereas implementation is the main barrier. The multi-agent workflow raises the share of scripts that run correctly as generated from 14.8% with a direct LLM call to 59.3%, reaching 80.6% on the four less complex problems, while tightly coupled intralogistics models remain an open challenge.

0.26

Physics-Aligned Electronic Ground-State Learning Improves Generalization

cs.LG physics.chem-phquant-ph ·2026-10-07 ·Eike S. Eberhard, Xaver Kainz, Viktor Kotsev, Abdulrahman Aldossary, Stephan Günnemann
Abstract

Machine-learned interatomic potentials (MLIPs) excel at in-distribution tasks, accelerating drug and material development, yet they struggle to generalize out-of-distribution. We propose to push the cost-accuracy Pareto frontier by designing observable-agnostic electronic ground-state descriptor models (GSMs) with computational costs situated between MLIPs and Kohn-Sham density functional theory (KS-DFT). We align the learning objectives and architectures of GSMs with the governing equations of KS-DFT by enforcing physical constraints and removing optimization pressure on unphysical or irrelevant degrees of freedom. In our size-extrapolation experiments from QM9 to QM40, our combined contributions OrthoNormal-Loss (ON-Loss) and Grassmann Restricted Occupied-Orbital Training (GROOT) reach a 79.1% energy and 83.4% force mean absolute error (MAE) reduction over previous state-of-the-art density GSMs. For Hamiltonian GSMs, ON-Loss and Residual Optimal-gauge Conditioning-aware KS-Eq. Training (ROCKET) together reduce the energy and force MAEs of the strongest baseline by 99.8% and 95.9%, respectively. Using a self-consistency rejection criterion, we filter out extrapolation errors on QMugs, rejecting fewer than 0.4% of predictions while reaching an energy MAE of 0.07 mHa. Finally, we demonstrate the efficiency of label-free self-consistency fine-tuning, and transfer GSMs to reactive chemistry in Transition1x, reaching energy errors below chemical accuracy.

0.26

An Informational Curse of Horizon in Goal-Conditioned Policy Learning

cs.LG cs.AIcs.RO ·2026-10-07 ·John L. Zhou, Yuxuan Dong, Jonathan C. Kao
Abstract · 25 pages, 11 figures

The difficulty of learning goal-reaching policies is often attributed to a "curse of horizon" that manifests as bias accumulation in temporal-difference backups and noisy advantage estimates. In this work, we identify an additional informational curse of horizon in goal-conditioned policy learning, where increasing the goal relabeling horizon can significantly reduce policy generalization and performance. Through a series of controlled experiments with oracle planners, we decouple the goal horizons sampled during training from those that the policy is asked to reach at test time. Even when evaluated only on a sequence of nearby subgoals, goal-conditioned behavioral cloning (BC) policies suffer from severe, training horizon-dependent performance degradation that is mitigated by reinforcement learning (RL) objectives. We explain this phenomenon as a horizon-dependent decrease in the conditional mutual information between actions and hindsight-relabeled goals, and find empirically that both BC and RL policies trained on longer-horizon goals exhibit a shift in sensitivity from goal to state information, as measured by the policy's input Jacobians. Motivated by this observation, we find that distilling the input Jacobians of short-horizon policies into long-horizon policies yields significant performance gains, especially in combinatorial manipulation tasks. Taken together, our results highlight goal relabeling horizon as an important consideration when learning generalist policies from offline data.

0.26

Evaluating Sequence Assembly Strategies for Differentially Private Synthetic Time-Series Forecasting

cs.LG ·2026-10-07 ·Guoxiong Long, Huizhen Huang, Qikun Cai, Tao Huang, Chen Hou
Abstract

Differentially private time-series generators commonly produce fixed-length synthetic windows, whereas downstream forecasting models often require long continuous training sequences. How these windows are assembled after generation can therefore alter the effective synthetic data presented to a forecaster, even when the trained generator remains unchanged. We study this post-generation sequence assembly process by systematically varying overlap rates and window-weighting schemes and evaluating the resulting sequences in terms of boundary continuity, statistical and temporal fidelity, and Train-on-Synthetic-Test-on-Real (TSTR) forecasting utility. Across four types of public datasets (ETTh1, ETTm1, Weather, and Appliances) and five forecasting models, the results reveal a clear forecaster-dependent assembly principle: downstream TSTR utility is jointly shaped by the forecaster, overlap rate, and window-weighting scheme, leading to distinct assembly preferences across forecasting models. Increased overlap generally improves boundary continuity, but improvements in continuity or individual fidelity diagnostics do not consistently reduce forecasting error, indicating that these diagnostics alone are insufficient for selecting assembly configurations. Complete five-forecaster assembly grids, together with matched Train-on-Real-Test-on-Real (TRTR) references, further characterize these regularities and quantify assembly-dependent utility relative to real-data training. We then validate the identified principles through additional analyses of robustness and generator variability.

0.26

Phase-HDC: Replacing Optimizer History with Gradient Thresholds in Discrete Phase Learning

cs.LG cs.AI ·2026-10-07 ·Ahmed Nebli
Abstract

Training a compact model often needs far more memory than storing it, because the optimizer keeps its own records of past gradients. For a hyperdimensional classifier whose learned parameters are low-bit angles, which we call a \emph{phase memory}, these records take several times more memory than the model itself. We ask whether such a model can be trained while storing nothing but the model. The proposed method, Phase-HDC, turns each stored angle by at most one step per update, against the sign of its current gradient, and only when that gradient is large enough. We show that this simple rule is the exact solution of a first-order loss model in which every changed parameter pays a fixed cost. When everything except the update rule is held fixed, Phase-HDC matches the accuracy of Adam with 6-bit moments while storing three times less. Across eleven image, tabular, and text datasets, it stores 16--23$\times$ less than standard float32 Adam and 4--6$\times$ less than 8-bit Adam. The price is an average loss of about five accuracy points against float32 Adam, while Phase-HDC is more accurate than 8-bit Adam on six of the eleven datasets, including byte-level text prediction, where 8-bit Adam collapses. Instrumented training runs explain these outcomes. Once parameters must sit on a discrete grid, Adam's moments mainly decide whether a parameter moves at all, a decision that a threshold on the current gradient can make without memory, and coarse quantization of the moments breaks this decision for inputs that the data rarely contain. The storage savings are logical state rather than measured hardware memory.

0.26

The AI Evaluation Ecosystem

cs.AI cs.CY ·2026-10-07 ·Yash Dave, Sang T. Truong, Serena Wang, Sanmi Koyejo
benchmarkposition
Abstract · 70 pages, 15 figures, 23 tables

AI evaluation shapes the decisions of model providers, users, funders, and regulators. We argue that designing valid benchmarks requires contextualizing design choices in the dynamics of this ecosystem of actors. We develop a simulation architecture that combines rule-based market dynamics with LLM-driven strategic actors, building on advances in Generative Agent-Based Modeling (GABM). We model benchmarks, consumer needs, and provider capabilities as vectors over a six-dimensional capability space (reasoning, coding, knowledge, safety, communication, agentic), with structural information partitions across actors. As a case study, we apply this stylized simulation to explore benchmark holdout design. We find that moving from public benchmarks to private holdout benchmarks shrinks the gap between benchmark scores and user satisfaction on most benchmarks but widens it on a few, depending on where holdout weights shift scoring credit. We stress-test our findings at both the instrument and case-study level, drawing on the V&V framework of Sargent (2013) and GABM-specific evidence criteria. Beyond holdout design, our simulation is a hypothesis-generating sandbox for studying how evaluator and policy choices, in turn, reshape the ecosystem.

0.26

Learning joint probabilistic weather forecasts from station observations alone

cs.LG stat.APphysics.ao-ph ·2026-10-07 ·Chaeyeon Yi, Yun Am Seo
Abstract · 62 pages, 6 figures, 3 Extended Data figures, 7 Extended Data tables; includes Supplementary Information

Assessing compound weather risks requires forecasts representing dependence between variables. CLARA (Calibrated Advection-Routing Attention) learns joint Gaussian predictive distributions of five surface variables from station observations alone, without numerical weather prediction or reanalysis; the approximately 28,000-parameter model supports CPU training and prediction. Across six multi-year folds on 96 stations, its lead-mean energy score is 4.9% lower than that of a learned comparator with matched temporal inputs (4.7% with a similar parameter count) and 11-65% lower than those of statistical baselines. Holding marginal variances fixed, removing learned correlations worsens joint negative log-likelihood by 1.0-2.8 nats per station. A covariance-scale estimator, proved consistent under stated assumptions, improves short-lead calibration but over-corrects at long leads. Synthetic interventions show an attention-bias coefficient alone does not measure forecast influence. Retrained in ten regions on six continents, CLARA outperforms persistence in all 60 multi-year region-lead comparisons and a similarly sized learned model in 57 of 60.

0.26

Gauss-Newton Accuracy and Indefinite Hessians: Uniform Coexistence in Low-Cost Sets

cs.LG stat.ML ·2026-10-07 ·Kihun Rhee, Hanjoon Byun
theory
Abstract · 28 pages, 1 table

We study the accuracy of Gauss-Newton curvature in ridge-regularized nonlinear least squares. Under local regularity and persistence of level-set curvature magnitude along an exact-fit section, we prove uniform coexistence of two curvature regimes. Global minimizers exist, and every global minimizer has relative Hessian error below $(1+\sqrt2)/8$, while the same low-cost set contains a point with an indefinite Hessian and relative error at least $15/8$. One positive ridge cap works for all independent center and label perturbations in fixed neighborhoods and every positive ridge weight up to the cap. These neighborhoods do not shrink as the ridge weight tends to zero. A pointwise certificate based on the current prediction level set controls the normal, mixed, and tangent parts of the Hessian correction. We prove a sharp relative-error bound over the stated pointwise class when the prediction map and ridge vary. Analytic examples describe the roles of output alignment, curvature orientation, and persistence. A separate structural result gives full Jacobian row rank throughout low-cost sets and exact interpolation near a rank-deficient reference.

0.26

DSReg: Provably Recovering Individual World Latents without Reconstruction

cs.LG cs.AIcs.ROstat.ML ·2026-10-07 ·Yujia Zheng, David Klindt, Randall Balestriero, Bernhard Schölkopf
theorycode/data
Abstract · Project page: https://dsreg.github.io/

Methods that recover individual latent variables of the world, from nonlinear ICA to dictionary learning and causal representation learning, anchor the latents to observations through reconstruction, auxiliary supervision, or distributional asymmetries such as non-Gaussianity. Methods without these anchors, including joint-embedding predictive architectures (JEPAs), identify the latent state only up to a linear transformation, so individual latents remain mixed. We close this gap: individual world latents can be provably recovered with no reconstruction, no decoder, and no labels. The key condition is Structural Diversity: different latents leave distinct dependency footprints on observations, just as no two snowflakes are alike. Building on the linear identifiability that LeJEPA provides, we prove that under Structural Diversity, DSReg (Dependency-Sparsity Regularization) recovers individual world latents up to signed permutation, without reconstruction or a decoder. It applies post hoc to any linearly identified representation, reusing trained checkpoints at no loss over joint training, and establishes the first fully identifiable JEPA that recovers every world latent. Moreover, as a condition on dependency footprints, Structural Diversity is strictly weaker than all structural conditions of prior identifiable latent variable models. Across synthetic regimes, world model probes, learned visual encoders, and external renderers, DSReg preserves dense prediction while improving individual-latent recovery and downstream use with scales.

0.26

How Many Repeated Pairwise Comparisons Are Needed for Ranking under Heterogeneity?

cs.LG cs.IT ·2026-10-07 ·Shashaank Aiyer, Han Shao
theory
Abstract · 53 pages, 4 figures

We study ranking models by population-average utility from pairwise comparisons when preferences vary across users and tasks. Prior work shows that a single comparison per user can be insufficient to identify the alternative with the highest average utility, even with arbitrarily many users (Golz et al., 2025). We investigate how many repeated comparisons within each user-task context are necessary and sufficient for ranking recovery. Under a heterogeneous Bradley-Terry model with fixed inverse temperature, we start with a naive MLE-based algorithm that requires $Ω(1/Δ^2)$ repeated comparisons per context to ensure ranking recovery. We then present two MLE-based variants and a randomized Russian Roulette-style algorithm that recover the ranking using $O(\log(1/Δ))$ repeated comparisons per context, and we prove that this logarithmic dependence is optimal. Despite this worst-case requirement, our Russian Roulette algorithm uses only $O(1)$ comparisons per context in expectation. Synthetic experiments and semi-synthetic experiments based on Arena data compare the four algorithms in settings with varying levels of preference heterogeneity and under varying context distributions.

0.26

Teaching PPG How not Who: Fixed-Effects Distillation from ECG

cs.LG ·2026-10-07 ·Zhongli Wu, Zhuangzhi Gao, Yuankai Wang, Gregory Y. H. Lip, Bilal H. Kirmani +1 more
unusual
Abstract

ECG is widely used to teach PPG-only models, yet what it teaches is unexamined. Wearables are valued for tracking how a person's cardiovascular state changes, but ECG-to-PPG distillation mostly learns who the person is. A per-recording mean, the trait, holds 40-59% of a frozen ECG teacher's target, and pooled students memorise it without carrying it to new recordings. The raw alignment cosine misses this, since a constant predictor scores 0.793. Across 34 runs, the more identity a student memorises, the less state it learns. Fixed-effects distillation subtracts each recording's mean from prediction and target, so the trait cancels exactly, while a pooled anchor keeps it. State agreement more than doubles, within-person labels improve while age and sex do not, and the gain holds on two backbones and two further databases. Conditioning on the recording turns distillation toward the within-person changes that wearables monitor.

0.26

Shape irregularity of Life-Like Network Automaton rules as an indicator of classification performance

cs.LG nlin.CG ·2026-10-07 ·Lucas C. S. Oliveira, Michiel Rollier, Jan Baetens, Odemir M. Bruno
unusual
Abstract

Complex Network (CN) classification requires high-level structural characterizations that are both scale-invariant and computationally efficient. Methods based on Life-Like Network Automata (LLNA) offer an interesting way to extract network descriptors by leveraging emergent temporal patterns without requiring provided features, but their efficacy is bottlenecked by a high-cost combinatorial optimization problem: the selection of the automaton transition rule. While current literature relies on exhaustive searches that are unfeasible for large-scale applications, this work reveals that the rule space is fundamentally structured by a property we term ``jaggedness'', that quantifies the resemblance of a LLNA transition function with a sawtooth shape. We demonstrate that this metric acts as a theoretical proxy for chaoticity and sensitivity -- properties essential for generating discriminative dynamic behaviors among network categories. Moreover, we introduce a heuristic search strategy that uses jaggedness to guide the rule selection. Experimental results show that our approach achieves classification accuracies within 5% of the global optimum while reducing computational overhead by 90% compared to exhaustive approach. Our findings provide a novel, efficient, framework for optimizing automata-based methods for pattern recognition.

0.26

Real-world application of deep learning in large-scale seismic interference attenuation: A case study in the Camie field of Angola

cs.AI physics.geo-ph ·2026-10-07 ·Jing Sun, Song Hou
published
Abstract

In marine seismic acquisition, seismic interference (SI) occurs when energy from nearby external seismic source(s) is captured. It typically appears as coherent noise with linear or non-linear movement and varying amplitudes across different sail lines. SI is commonly observed and poses a challenge for seismic data processing. We present a case history of a previously proposed deep neural network (DNN)-based workflow applied for SI attenuation across a marine seismic block in the Camie Field of Angola. This field survey covers over 345 km2 and is marked by the challenge of multiple SI types. The employed DNN-based workflow performs SI attenuation in the common shot domain based on a supervised learning framework: a small subset of the SI-contaminated data was first processed by a conventional geophysical algorithm to obtain an estimate of the SI noise, which was then manually blended with the SI-free common shot gathers from the same survey to generate the training pairs. To ensure signal fidelity, several techniques were applied to improve the DNN's performance. A key highlight of this case history is its scale: this represents a real-world, large-scale processing project and we present a comprehensive comparison of the DNN-based workflow with the conventional geophysical algorithm across the entire survey block, focusing on both processing quality and processing time. The results demonstrate the outstanding performance of the employed DNN-based workflow, which achieved higher SI removal accuracy, with less signal leakage and more complete SI removal. The promising results of this application also open up possibilities for integrating deep learning into other seismic denoising tasks. In addition, we discuss the limitations of this case history, aiming to provide insights for future research and applications in the field.

0.26

Controlling Dependence in Implicit Generative Models via Spread Mutual Information

cs.LG stat.ML ·2026-10-07 ·Jiahao Yu, Song Liu, José Miguel Hernández-Lobato, RuiKang OuYang
unusual
Abstract

Mutual information (MI) provides an objective for suppressing or encouraging statistical dependence in implicit generative models. However, direct MI evaluation is challenging in implicit models due to typically intractable densities. A remedy is estimating the generator gradient from the difference between conditional and marginal scores. This score difference can, in turn, be estimated by differentiating a log density ratio learned through classification. This construction nevertheless faces two difficulties: (i) singular distributions need not admit the required score functions, and (ii) poor overlap can hinder density-ratio estimation. We therefore introduce Spread Mutual Information (SMI), a weighted integral of MI across noise levels obtained by applying a common spreading kernel to the generated variable. Gaussian spreading yields smooth, strictly positive conditional and marginal densities, extending the gradient construction to distributions that may originally be singular. Across a variaty of experiments, SMI consistently achieves effective dependence control among MI-based methods and remains competitive with established task-specific approaches.

0.26

From Plausible Hierarchies to Useful Taxonomies: Evaluating Agentic Harnesses on Customer Feedback

cs.AI cs.LGcs.SC ·2026-10-07 ·Prabhath Chellingi, Raviraja G, Viraj Bagal
unusual
Abstract

Taxonomies are the symbolic representations through which AI systems organize evidence, aggregate patterns, and answer questions over large document collections. Over customer feedback, the category tree decides how every record is counted and routed, which problems get seen, and which team owns them. Agentic harnesses now make it easy to generate a plausible-looking hierarchy, and such trees are checked today with generic, individually scoped checks: each name fits its description, sits under the right parent, and stays distinct from its siblings. We ask a more operational question: when is a generated hierarchy actually useful as a production taxonomy? We build six taxonomies over two proprietary feedback corpora (1,940 and 5,000 records): for each corpus, a production reference and two repeated runs of the same harness under identical inputs. All six pass every generic naming and structure check, and a deeper product-coverage check even prefers the generated trees. Yet in every generated tree at least 97.7% of leaf names merely restate an ancestor's name (13.9% and 2.9% in the references), and in one, three of every four records fall under multiple top-level categories. We introduce two families of whole-tree metrics: structural discriminators test whether a tree's shape was learned from the data or imposed by its generator; team partitionability tests whether branches split feedback into groups teams can own. Trees the generic checks rate as equally correct differ by 27 percentage points in cross-branch leakage, and only one of the two beats a random split. Surface plausibility is an insufficient measure of taxonomy quality: evaluation must measure the whole tree as well as each node.

0.26

Logarithmic Regret via Passive Change Detection in Piecewise-Stationary Self-Tuning Regulation

cs.LG cs.AI ·2026-10-07 ·A. Ch. Madhusudanarao, Rahul Singh
theory
Abstract

We study minimum-variance control of an unknown autoregressive system with exogenous inputs and coefficients that change at unknown times. Under bounded independent disturbances, fixed detection gaps, stability and feasibility conditions, and sufficient time between changes, we prove \(O((C+1)\log((T+1)/δ))\) regret with probability at least \(1-δ\), where \(T\) is the horizon and \(C\) the number of changes. Unlike switching bandits, where unselected arms can change unobserved, admissible plant changes provide information during exploitation: the correct feasible controller leaves only the disturbance in the output, whereas a detectable change raises output energy under the old controller. PIECE-CD explores initially and after alarms, then uses gated recursive least squares for control. Its energy test compares windowed output power with a threshold above the noise floor; the extension to unstable controller mismatches also monitors the reference controller's input proposal. We control false alarms across the horizon and prove logarithmic detection delay. Inputs are clipped to prescribed bounds. Logarithmic regret also holds under an explicit condition ensuring that clipping becomes inactive after a finite burn-in. Under the stated feasibility conditions, the extended detector covers destabilizing changes with detectable excess energy over a fixed window.

0.26

Koopman Observers for Diffusion Acceleration: Correcting Feature Forecasts with Shallow Measurements

cs.LG math.DSstat.ML ·2026-10-07 ·Hanru Bai, Yuanchao Xu, Fengyi Li
Abstract · 14 pages

Feature caching accelerates diffusion sampling by replacing expensive network evaluations with predictions from previously computed activations. However, forecasts based only on past features cannot directly incorporate changes in the current denoising state. We investigate whether inexpensive, freshly computed features can serve as observations for correcting these predictions. We introduce an observation-corrected Koopman framework for accelerating frozen diffusion models. Using calibration trajectories, we identify finite-dimensional, time-dependent Koopman approximations that jointly describe the increments of shallow and deep network features. During accelerated sampling, these operators predict the evolution of expensive deep features, while innovations in the observed shallow features correct the predicted state. Periodic full evaluations refresh the observer, and all generative-model parameters remain unchanged. This formulation enables controlled comparisons of temporal prediction and observation correction. Across three 10,000-image runs per dataset, our method reduces paired Inception-feature MSE by $19.9\%$ on CIFAR-10 and $11.9\%$ on a ten-class ImageNet subset relative to channelwise affine prediction under the same four-partial-step schedule. Matched ablations attribute additional reductions of $4.54\%$ and $4.67\%$ to observation correction. The observer achieves $1.89\times$ and $1.85\times$ measured speedups over DDIM-50, supporting improved reference-sampler fidelity without retraining the denoiser.

0.26

What Can a Gaussian Process Design Test

cs.LG ·2026-10-07 ·Ivan De Boi, Marnix Van Soom
unusual
Abstract

A Gaussian process (GP) model can agree with the data for two reasons: its assumptions are right, or the chosen inputs could never have shown that they are wrong. The distinction can be checked from the design before any responses are observed. Every model implies relations that its noiseless responses must satisfy at the chosen inputs, such as the middle value lies on the line through its two neighbours. For GPs built from finitely many features, these relations are exactly the null space of the kernel matrix. Gale duality gives them a geometric interpretation, in which each observation has a vector and the smallest groups of observations that can expose an error are the circuits. For other kernels the relations become soft: response patterns may be improbable under the prior rather than algebraically impossible. A standard test then combines two kinds of evidence. Structural evidence comes from a violated relation and grows without limit as the noise falls. Prior-based evidence only says that a departure is improbable under the prior. With all inputs at the two ends of an interval, for example, a GP can reject a straight line against a large curvature, but only because the implied intercept is improbable, never because curvature was seen. In simulations the predicted power matched the observed rejection rates. Choosing the next input by predicted power raised the power against a localised discrepancy from 0.48 to 0.72, against 0.51 when choosing by predictive variance, and a grid in two dimensions contained exact tests of additivity that a Latin hypercube lacked. The test itself is classical. The contribution is the prospective reading of that test: before observing the responses, the design already determines what kind of contradiction it can produce.

0.26

A Closed-Loop Non-Asymptotic Convergence Analysis of PPO with Learned Critics and Clipping

cs.LG ·2026-10-07 ·Junwei Su, Mengfan Liu, Yanyong Zhang, Chuan Wu
Abstract

Despite its widespread use, Proximal Policy Optimization with clipping (PPO-Clip) remains difficult to tune, and the interactions among critic learning, clipping, and rollout reuse remain incompletely understood. We develop a \emph{non-asymptotic} analysis of PPO-Clip as a \emph{closed-loop actor--critic} system. It captures actor--critic coupling, nonsmooth probability-ratio clipping, finite-batch reuse, and predictable early stopping under explicit coverage and critic regularity assumptions, using raw GAE and Monte Carlo critic targets. Our synchronous and asynchronous guarantees jointly characterize policy stationarity and the tracking accuracy of the learned critic, with explicit dependence on algorithmic parameters. A sufficient coupling condition gives optimization, critic tracking, clipping, and finite-batch errors a common amplification bound. The asynchronous result also requires a delay-dependent critic stepsize restriction; violating these conditions does not establish divergence. For finite layered MDPs with tabular critics, a uniform bound on the actual clipped-gradient class replaces complete-trajectory counting. A verified growing-horizon family has polynomial sample complexity, and a two-time-scale schedule gives $O(T^{-2/5})$ stationarity and critic-tracking bounds with explicit fresh-rollout accounting. These results together advance our understanding about PPO and provide theoretical guidance in tuning.

0.26

CircuitTree: Composing Circuits into A Syllable Transcriber

cs.LG cs.AIcs.SD ·2026-10-07 ·Jiankun Wei, Ewan Dunbar, Gerald Penn
unusual
Abstract

The goal of circuit discovery is to decompose a monolithic neural network into tiny modular subnetworks which are interpretable in the sense that each does an identifiable subtask. We develop a practical method for composing arbitrarily many circuits into a single model. We propose a greedy, density-scaled composition algorithm that overcomes limitations of previous attempts to compose circuits by naive union, then generalize it to $N$ circuits via CircuitTree, a new data structure built on a $Θ(N\log N)$ tree-compose schedule. We compose circuits pruned from HuBERT for classifying individual phonemes into a syllable classifier, a capability none of the 39 phoneme circuits has individually. Chaining CircuitTrees into a hierarchy-of-circuits pipeline, we transcribe open syllables at 64.8\% accuracy, within 1.2 points of a 66.0\% ceiling that shares no circuitry at all, with 99.5\% of circuit components shared.

0.26

Kuration SDK: Addressing the Virtual2Real Gap via Data Curation

cs.LG cs.AIcs.CVcs.RO ·2026-10-07 ·Nirmit Desai, Eric Song, Mayank Sengupta, Tejal Bedmutha, Siri Reddy +2 more
code/data
Abstract

Benchmarks for measuring the quality of action-conditioned world models are still evolving and shifting away from visual similarity-based metrics to action-semantic and physically-grounded metrics. However, for domain and task-agnostic action-conditioned world model training, existing benchmarks provide a limited signal. By training and evaluating diffusion world models on CounterStrike gameplay data, we confirm that qualitative playability does not correspond with metrics such as FVD, LPIPS, and JEDi. We term this the Virtual2Real gap. We posit that, in lieu of reliable benchmarks, curating raw gameplay data and measuring a variety of diagnostic properties provides a more robust signal to bridge the gap, before the training even begins. We present several curation strategies and a general-purpose kit for physical AI data curation called Kuration SDK, which is being open-sourced with this paper. The SDK was instrumental in uncovering the root cause of the virtual2real gap in a specific case: why two world models trained on identical gameplay map, action and state distribution, behaved very differently when played in spite of having very similar LPIPS and FVD scores. Thus, Kuration SDK has the potential to uncover the root causes of Virtual2Real gap in specific datasets and accelerate development of sample-efficient training datasets.

0.26

Controlled Acquisition and Abstention in Three-Channel Score Conflicts

cs.LG cs.MMeess.AS ·2026-10-07 ·Mengzhe Geng
benchmark
Abstract

When audio, video, and text disagree, accuracy alone does not show whether to acquire another source or abstain. We study these choices in a controlled three-score benchmark: a policy observes two signed scores, may request the third at a cost, and can abstain. The primary reward is mechanism-specific: abstention is correct only for one designated ambiguity mechanism and is penalized under mixed corruption. Matched controls show that a threshold policy matches always-request decisions with fewer requests; its advantage over always-answer fusion depends on the reward assigned to that ambiguity. On a partially held-out synthetic split, the threshold policy reaches 0.789 +/- 0.006 targeted decision accuracy and 0.481 +/- 0.014 utility across 83 seeds. A three-score majority reference reaches 0.626 +/- 0.008 and 0.252 +/- 0.016, but uses more information. In a matched-budget test, a train-only value selector improves utility over no-query and matched-random policies at 10% and 25% budgets, while pair uncertainty has higher utility at every budget. At 50% and 63.7% budgets, the selector lowers utility despite slightly higher non-ambiguous accuracy. If all abstentions are scored incorrect, majority outranks the threshold policy in utility. At a central temporal setting, full-trace controls match the neural models while position perturbations separate them. On held-out-actor emotion clips, eight-frame fusion has opposite-signed accuracy differences for two encoder pairs, with both actor intervals containing zero; matched-request routing gains are small and uncertain. These results separate full-modality accuracy from pre-request selection value and show that selection value depends on budget and the observed-pair ranking.

0.26

Decoupled Optimization for Teacher-Student Semi-Supervised Learning via a Pioneer Student

cs.LG ·2026-10-07 ·Haorong Han, Jidong Yuan, Chixuan Wei, Yongqi Sun
Abstract

Semi-supervised learning (SSL) relies on two core mechanisms: self-training under the Teacher-Student (T-S) framework and joint optimization of labeled and unlabeled losses. Despite their effectiveness, we find both mechanisms introduce distinct optimization pathologies. First, parameter coupling enforces strict synchronization between teacher and student, where strong regularization on the student degrades the teacher's fitting ability, thereby limiting the permissible generalization intensity. Second, the imbalance in gradient update consistency between labeled and unlabeled losses drives the shared parameters to prematurely converge to labeled-dominated local minima, creating a bottleneck for global optimization. To address both issues, we propose the Pioneer Student (PiS), an auxiliary branch that operates in an independent parameter space and periodically transfers accumulated knowledge back to the T-S model. Extensive experiments show that PiS is a universal plug-and-play module that consistently improves mainstream SSL methods.

0.26

OrthoGen: A Generative Orthogonal Learner for Time-Varying Treatments

cs.LG ·2026-10-07 ·Tomàs Garriga, Valentyn Melnychuk, Konstantin Hess, Eduard Serrahima de Cambra, Axel Brando +2 more
Abstract

Estimating conditional distributional potential outcomes (CDPOs) over time is important in medicine (e.g., to estimate patient-specific risks under different treatment sequences). However, this task is challenging because of time-varying confounding, yet existing adjustment strategies for this task are limited. In this paper, we aim to learn CDPOs under time-varying treatments using flexible generative models. Our contributions are two-fold. (1) We introduce a tailored adjustment strategy for our setting, namely, generative recursive g-computation. Our adjustment strategy recursively propagates full conditional outcome distributions rather than conditional means, modeling the variables of interest directly rather than full trajectories. Building on our adjustment strategy, we formulate simple generative learners for CDPO estimation. However, these learners can be sensitive to nuisance estimation errors, which motivates an orthogonal learner. (2) We thus introduce OrthoGen, a Neyman-orthogonal and doubly robust generative learner. Importantly, we show that OrthoGen further achieves rate double robustness and quasi-oracle efficiency under suitable conditions. Our learners are flexible and can be instantiated with different generative backbones (e.g., normalizing flows and diffusion models). Across experiments with synthetic, semi-synthetic and real-world datasets, we find that OrthoGen is highly effective. To the best of our knowledge, we are the first to propose a generative orthogonal learner for estimating CDPOs under time-varying treatments.

0.25

Distilling Graph Geometry: Knowledge Gap from GNNs to MLPs

cs.LG ·2026-10-07 ·Zhewei Chen, Hao Zhu, Jiaojiao Jiang, Ahad N. Zehmakan
Abstract

GNN-to-MLP distillation aims to retain the predictive accuracy of a message-passing teacher while deploying a graph-free MLP at inference. Existing methods mainly transfer node-wise predictions or use confidence-based reweighting, but they do not specify where the student should preserve the teacher's graph-induced geometry. We show that this omission leads to two spectral failure modes in the student's representation space. On sparse graphs, the student suffers from spectral underfit, missing high-energy teacher directions induced by aggregation. On dense graphs, it suffers from spectral overfit, retaining spurious directions that the teacher has collapsed through aggregation. Motivated by an energy-weighted teacher-student alignment objective, we propose Graph Geometry-aware MLP (G^2MLP), a training-time distillation framework guided by Ollivier-Ricci curvature. Curvature acts as a local proxy for allocating supervision between prediction-level and representation-level alignment. The deployed model is a standard MLP with no graph access at inference. Across node-classification benchmarks, G^2MLP improves over graph-free distillation baselines in most settings, reduces the teacher-student rank gap in both regimes, and transfers to Graph Transformer teachers and link prediction.

0.25

UniCSI Towards a Universal Wi-Fi CSI Encoder for Ubiquitous Human Sensing

cs.LG cs.NI ·2026-10-07 ·Daniel Eckhoff, Hua Kang, Zhitang Chen, Jie Chuai
Abstract

Wi-Fi sensing promises to turn the everyday wireless signals that already surround us into ubiquitous sensors for human sensing. However, a fundamental obstacle is that CSI is acquired under diverse device-specific configurations, including different subcarrier counts, bandwidths, and carrier bands. Consequently, the resulting CSI tensors vary in both spectral resolution and tensor shape, making heterogeneous modeling challenging. Standard architectures struggle with such heterogeneity, forcing lossy pre-processing which compromises the underlying signal. To bridge this gap, we present UniCSI, a unified foundation architecture that directly operates on heterogeneous CSI while preserving the integrity of the native waveform. UniCSI hinges on two core innovations: (1) a physics-informed RF tokenizer that encodes each frequency channel based on its fractional position within the physical spectrum rather than rigid array indices. It preserves intrinsic spectral coherence and enables seamless, frequency resolution-agnostic processing across arbitrary sensing configurations. (2) a spectral aggregator that distills variable-length channel sequences into a fixed-size spectral signature, effectively decoupling the feature dimensionality from the physical subcarrier spacing. Extensive evaluations on a large-scale corpus of 25 heterogeneous public datasets, spanning 14 to 2048 subcarriers, 20 to 160 MHz bandwidth, and the 2.4 and 5 GHz bands, demonstrate that native heterogeneous ingestion substantially improves cross-domain transfer under both supervised and self-supervised training schemes, particularly in regimes where fixed-grid architectures fail to generalize.

0.25

What the Sleeve Feels: Explainable Machine Learning for Textile Pressure-Based Postural Screening

cs.AI ·2026-10-07 ·Limon Bin Hossain, Md Sadib Rahman Ananta
Abstract

Pressure-sensing smart textiles convert body-surface contact into a dense, image-like signal closely tied to posture and movement, making them a promising low-cost route to wearable posture screening. Realizing that promise, however, requires more than classification accuracy: a deployable system must generalize to wearers unseen during training, expose the physical evidence behind its decisions, and tolerate the small donning offsets that occur whenever a garment is removed and re-worn. This paper addresses these three requirements jointly using a knitted piezoresistive sleeve worn on the forearm as a testbed. We regroup fine-grained everyday activities into three coarser screening categories (neutral, potentially undesirable, and functional or transitional), engineer 29 interpretable pressure-distribution features spanning global intensity, spatial center of pressure, quadrant asymmetry, distribution complexity, and short-horizon temporal change, and evaluate under a strict subject-wise split. A tuned XGBoost classifier reaches 0.818 accuracy, 0.788 balanced accuracy, and 0.801 macro F1 on unseen test subjects, with tight frame-level bootstrap 95% intervals of about plus-minus 0.01 and a subject-to-subject standard deviation near 0.06 under leave-one-subject-out cross-validation. A simple 2D-CNN baseline trained on raw frames achieves broadly similar performance, showing that hand-engineered features are not left behind by a learned spatial representation on this task. SHAP-based explanation, a feature-group ablation, per-activity error analysis inside the pooled undesirable class, class-mapping sensitivity, and a simulated donning-rotation stress test together locate what the model relies on, where it degrades, and why, directly targeting the generalization, interpretability, and robustness gaps that determine whether such a system is deployable.

0.25

Efficient Patch-Based Anomaly Detection Fused with Diffusion Driven Generative Modeling for Semiconductor Wafer Bin Map Open Set Anomaly Detection

cs.LG cs.AI ·2026-10-07 ·Limon Bin Hossain, Md Sadib Rahman Ananta
benchmark
Abstract

Spatial defect signatures on wafer bin maps (WBMs) trace yield loss to specific process faults, yet supervised classifiers recognize only the defect types seen during training, and one-class detectors built on a single mechanism tend to capture either local structural deviations or global distributional violations, but rarely both. This work proposes a hybrid one-class framework that couples a patch-based student-teacher detector (EfficientAD) with a denoising diffusion probabilistic model (DDPM) used for partial-diffusion reconstruction, and fuses their percentile-calibrated scores through a fixed convex combination. Trained on only 700 normal wafers from the WM-38K mixed-type dataset and evaluated on 18,658 held-out wafers, the fused detector reached an AUROC of 0.9985 and reduced misclassifications from 852 (DDPM) and 1,412 (EfficientAD) to 618, with all pairwise differences significant at p < 0.001. Beyond aggregate accuracy, the analysis shows that the gain arises from weakly overlapping errors between the two modules, yet fixed-weight fusion recovers only 40-70% of the correction available to an oracle selector. Under the benchmark's inverted class balance, average precision and F1 saturate, while the Matthews correlation coefficient and negative predictive value expose unreliable normal predictions. Pixel-level maps further show that strong image-level separability does not imply spatial localization, and the diffusion module succeeds as a local density prior rather than through global geometric reasoning. These findings motivate sample-adaptive fusion and imbalance-aware evaluation of hybrid wafer anomaly detectors.

0.25

How Narrative Wrapping Affects LLM Refusal: A Cross-Language Benchmark and Defense

cs.AI ·2026-10-07 ·Zhankai Ye, Yanning Wang, Yukai Jin, Bo Mei, Fangyi Li +3 more
benchmark
Abstract

Safety-aligned language models often refuse a harmful request stated directly but answer the same request inside a role-play or narrative wrapper. We measure this vulnerability across languages and registers: attack success on Qwen3-1.7B is already 89.4% in English and 93.0% in modern Chinese, and reaches 95.7% in Classical Chinese. We build GUISE, a benchmark for systematically studying this vulnerability. It includes parallel requests in English, modern Chinese, and Classical Chinese, matched harmful and benign pairs, wrapper types held out for evaluation, and a stricter criterion that counts warn-then-answer responses as attack successes. Representation analysis shows that language and register move harmful-request representations only slightly away from the model's refusal direction, whereas narrative wrappers move them much farther away. We propose AXIS, which combines preference optimisation with a rotation objective that aligns harmful-request representations with the refusal direction and a commitment objective that trains the model to refuse completely rather than produce a warn-then-answer response. Across Qwen3-1.7B, Qwen3-4B and GLM-4-9B, AXIS achieves the highest combined safety and usability score among the compared methods.

0.25

miniTE: Diagnosing and Bridging the Execution Gap in Native NVFP4 Reinforcement Learning

cs.LG cs.AIstat.ML ·2026-10-07 ·Zhen Li, Mingfa Feng, Yiming Zhang, Yang Yu, Changhai Zhou +4 more
benchmark
Abstract

Native NVFP4 promises to accelerate reinforcement learning (RL) for large language models (LLMs), but achieving stable learning across separate training and sampling engines remains challenging. We show that matching model weights and numerical formats is insufficient to ensure consistent policy evaluation, highlighting the importance of cross-engine numerical execution. We introduce miniTE, a lightweight framework that combines execution diagnostics, cross-engine numerical alignment, and efficient native W4A4 (4-bit weights and activations) GEMM computation for LLM RL. It improves numerical execution consistency without modifying the RL objective, while reducing implementation overhead to realize the benefits of native low-bit computation. Long-horizon RL experiments on Qwen3-30B-A3B demonstrate stable training and benchmark performance close to BF16, outperforming naive Transformer Engine (TE) NVFP4 in evaluation quality. On single NVIDIA B300 GPU node, the optimized learner reaches 33,540 tokens/s, yielding 1.39x and 1.47x the throughput of BF16 and TE NVFP4, respectively, for complete learner updates. We released the framework, kernels, diagnostics, and training configurations to make native low-bit RL easier.

0.25

Gaussian Equivalence for Multi-Head Self-Attention

cs.LG stat.MLmath.PR ·2026-10-07 ·Tomohiro Hayase, Ryo Karakida
Abstract

A theoretical understanding of multi-head self-attention is fundamental to the study of modern neural networks. Using random matrix theory, we establish Gaussian equivalence for multi-head self-attention: replacing softmax attention with rescaled scores plus Gaussian noise preserves the limiting spectral law of the centered output. This equivalence also covers value and output projections that depend on the keys. The resulting laws separate the effects of head allocation and projection widths, and distinguish spectrum-preserving across-head sharing from within-head key--value dependence.

0.25

The Identifiability and Observability of Deep Normalized Attention

cs.LG ·2026-10-07 ·Pranav Venkata Konda
Abstract

We study which parameters of deep, unmasked, single-head attention are determined by its input--output function. For known positive nonconstant real-analytic normalizers, the function generically determines the effective scores and combined value map up to the signs induced by even normalizers. This proves the real-analytic case of a conjecture of Henry--Marchetti--Kohn, including softmax. We then classify exceptional fibers under explicit normalizer conditions, identifying when collapse makes later scores unobservable, and establish sharp Taylor orders for local identification. Near simultaneous query/key collapse, we compute the complete native Jacobian decay spectrum on separating finite input banks. For common first nonconstant normalizer degree $k$, layer $i$ has contact order $2k3^{i-1}-1$, with exact multiplicities and kernel dimension. High-precision and automatic differentiation calculations illustrate the resulting loss of numerical sensitivity.

0.25

Self-Evolve With a Reference:Anchored Training of Tool-Integrated Agents

cs.AI ·2026-10-07 ·Wenjie Liao, Liangjie Zhao, Zehong Cao
Abstract

Self-evolving tool-integrated agents learn from tasks and feedback generated within their own training loop. A Curriculum Agent generates tasks, while an Executor Agent learns from self-consistency signals through reinforcement learning. However, relying solely on the current Executor for feedback has two limitations: group-relative advantages vanish under full consensus, while uncertainty-based curriculum rewards favor disagreement without showing whether the generated tasks support further learning. These limitations motivate an additional reference beyond the current Executor. We propose \textit{AnchorLoop}, which introduces a frozen copy of the previous iteration's Executor as a historical reference and reuses it on both sides of the training loop. For the Executor, the anchor provides a cross-reference advantage that evaluates current outputs against both current and historical majority answers. For the Curriculum, it provides an agreement-based reference based on differences in sampled majority agreement. Since the Executor and anchor have identical parameters during Curriculum training, this comparison serves as a proxy for task selection rather than evidence of inter-version improvement or correctness. Across 13 reasoning benchmarks, AnchorLoop improves over Agent0 by 2.5\% on mathematical reasoning and 2.8\% on general reasoning tasks. It also maintains higher effective-advantage variance and continues improving in later iterations as the unanchored baseline shows diminishing gains. These results demonstrate the benefit of introducing a lightweight historical reference into self-evolving tool-integrated agents without external task or answer supervision.

0.25

JevForest: Path Voting for Budgeted Feature Acquisition

cs.LG stat.ML ·2026-10-07 ·Yu Yan
Abstract

Choosing which information to observe is central to prediction under limited observation budgets. We study JevForest, a feature acquisition policy that aggregates path-dependent proposals from bootstrapped trees, weights them by global training information gain, and predicts from the acquired values with a shared masked classifier. An online implementation queries Jev for semantic answers selected by this policy. On small balanced held-out samples, four-question forest acquisition achieves accuracy $0.729$ on AG News ($n=48$), compared with $0.667$ for a static gain ranking and $0.583$ for random ordering. On TREC ($n=24$), the ordering reverses: forest accuracy is $0.667$, compared with $0.750$ and $0.833$. Asking all eight questions in one batch yields higher accuracy at lower measured cost and latency than four sequential forest queries; direct Jev classification matches the batch accuracy while costing less. Offline MiniBooNE experiments yield accuracy $0.845\pm0.010$ at ten features and $0.885\pm0.008$ at forty features over three jointly varying data and forest seeds (mean $\pm$ sample standard deviation). A companion Newton boosting implementation provides preliminary full-feature synthetic results. These exploratory findings establish a working Jev acquisition workflow but do not support a general advantage for path voting: its value depends on the task, predictor, and the distinction between question budgets and actual query costs.

0.25

Dual- versus Single-Suggestion AI Support for Radiographic Interpretation in Residents: Randomized Multireader Study

cs.AI cs.CY ·2026-10-07 ·Lin Wu, Zhe Xu, Hongyi Wang, Feifei Zhou, Wei Deng +8 more
Abstract

Purpose: To compare dual- and single-suggestion AI support for radiographic interpretation by residents, particularly when the shared AI suggestion was incorrect. Materials and Methods: This prospective, multicenter, randomized three-arm reader study was conducted at three hospitals in China from July to September 2026 (ChiCTR2600129243). After specialty stratification, 132 residents with fewer than 3 years of clinical experience were randomized 1:1:1 to GPT-5.4 alone (group A), GPT-5.4 plus Kimi-K2.6 (group B), or GPT-5.4 plus Gemini-3.6 Flash (group C); 123 were analyzed. Participants interpreted 60 radiographs before and after AI support. The primary outcome was accuracy change. Welch ANOVA and Holm-adjusted t tests compared support conditions; HC3 linear models assessed specialty interaction. Results: Among 123 residents (mean age, 24.1 years +/- 1.4; 65 women), radiology residents showed greater accuracy improvement with dual- than single-suggestion support (B-A, 6.69 percentage points [95% CI, 0.97-12.40]; C-A, 7.87 percentage points [95% CI, 1.64-14.11]; Holm-adjusted P = .030 for both), whereas accuracy change did not differ in non-radiology residents (P = .20). When GPT-5.4 was incorrect, AI-assisted accuracy was higher with dual- than single-suggestion support in radiology residents (40.1% and 40.4% vs 20.0%) and non-radiology residents (31.3% and 31.0% vs 12.1%) (all Holm-adjusted P < .001). The dual-suggestion effect differed by specialty (interaction difference, 10.44 percentage points; 95% CI, 4.36-16.52; P < .001). Conclusion: Dual-suggestion support may mitigate the influence of erroneous AI suggestions, with greater accuracy improvement observed in radiology but not non-radiology residents.

0.25

Think Before You Paint: Recursive Latent Reasoning for Diffusion Models

cs.AI cs.LG ·2026-10-07 ·Paweł Skierś, Małgorzata Grzanka, Wojciech Masarczyk, Jan-Willem van de Meent, Kamil Deja
Abstract

Diffusion models generate realistic images but often fail on visual reasoning tasks, such as filling in a Sudoku or drawing the path through a maze. When a discrete symbolic representation is available, recursive methods such as the Tiny Recursive Model (TRM) solve even hard instances of these puzzles. We ask how such reasoning can be carried over to pixels, where no symbolic representation is available. We propose Painter-Thinker (PaTh): a small recursive network (the Thinker) reasons over a grid of learned tokens that encode the noisy image and the conditioning, refines a latent state within every denoising step, and steers a frozen diffusion model (the Painter) through ControlNet adapters. The Thinker is trained with the standard reconstruction loss alone, without symbolic targets, a solver, or a verifier. PaTh solves 92.5% of hard MNIST Sudoku puzzles (prior best 75%) and 71.2% of extreme ones (prior best 4.1%), with 10M parameters against 82M for a standard diffusion model. It also improves on mazes, Queens, and CLEVR scenes with specified spatial relations, and its advantage grows with problem size. Diagnostic experiments show that PaTh recovers from injected mistakes that the diffusion model cannot repair, especially when many cells are wrong. Together, these results show that reasoning mechanisms developed for symbolic data can be integrated into pixel-space diffusion without symbolic supervision, opening a path toward generating data under increasingly complex constraints.

0.24

Dataset Pruning from First Principles: A Label-Free Linear Programming Approach

cs.LG stat.ML ·2026-10-07 ·Rodrigo Schuller, Francisco Ganacim
Abstract

Dataset pruning reduces a large training set to a representative subset while preserving model performance. Existing geometry-based methods typically assume that nearby points in embedding space share similar properties. Rather than imposing this assumption, we derive geometric selection criteria by reformulating unbiased subset selection as a variance minimization problem. Unbiasedness ensures that unweighted subset averages recover full-dataset averages in expectation, including losses and gradients at fixed model parameters. Specifically, we characterize a family of unbiased subset selection algorithms as a high-dimensional polytope. In this context, minimizing the expected sampling variance is a linear objective. Differences in sampling variance, averaged over rigid motions, admit closed-form pairwise expressions. Because the polytope has high dimension, directly applying standard linear programming is impractical. We instead use these expressions to construct an efficient vertex walk that optimizes an approximation of the variance objective while preserving unbiasedness, yielding a method that requires neither labels nor model training during selection. Across CIFAR-10, MNIST, and CelebA benchmarks, our method matches or exceeds uniform sampling in mean test accuracy at every evaluated budget and outperforms competing geometric methods in several settings, particularly at small selection budgets. Beyond dataset pruning, the same framework reduces stochastic-gradient variance by increasing diversity within mini-batches while keeping the batch size unchanged.

0.24

Beyond Policy Support: Interaction Constrained Offline Reinforcement Learning for Autonomous Driving

cs.LG cs.AIcs.RO ·2026-10-07 ·Mahmoud Selim, Cristina Cipriani, Karl Henrik Johansson
Abstract · Corrected affiliation formatting and some typos

Offline reinforcement learning enables reward-driven policy improvement from fixed datasets without requiring online exploration, making it particularly attractive in safety-critical domains. A central challenge, however, is distribution shift: policy optimization may favor actions that are weakly supported by the offline data, rendering value estimates unreliable. Existing approaches primarily control this shift in the policy's own action space. In interactive environments such as autonomous driving, this can be insufficient: a candidate ego trajectory may remain well supported under the marginal behavior distribution while being poorly supported jointly with the surrounding-agent behavior observed in the logged interaction. We refer to this degradation in interaction support as \emph{interaction distribution shift} (IDS), and introduce \emph{Interaction-Constrained Drive Policy} (ICDP), an offline reinforcement learning framework that explicitly controls interaction-level distribution shift. Starting from the joint data distribution over ego and surrounding-agent futures, we show that joint-support degradation decomposes exactly into an ego-support component and a residual interaction-support component. We recover the latter through contrastive density-ratio estimation, isolating interaction compatibility without explicit joint-density modeling, surrounding-agent prediction, or rollouts in reactive simulators or learned world models during policy optimization. Closed-loop evaluations on nuPlan, Interplan and real-world truck experiments show that ICDP suppresses high-value yet interaction-unsupported trajectory selections and improves performance in interaction-critical driving scenarios. Project webpage: https://mahmoud-selim.github.io/ICDP/

0.24

Composing What Each Teacher Learned: Delta Distillation for Multi-Teacher On-Policy Learning

cs.LG cs.AI ·2026-10-07 ·Yuanda Xu, Hejian Sang, Zhengze Zhou, Shayan Mohajer Hamidi, Xiaomin Li +2 more
benchmarkcode/data
Abstract

Multi-teacher on-policy distillation (MOPD) is used in two settings. In common-domain composition, several teachers score each student rollout from one prompt domain and their signals form a single target; in routed-domain distillation, prompts from different domains are assigned to the corresponding specialist. Both settings usually transfer each teacher's endpoint policy, which mixes what post-training changed with preferences inherited from the teacher's base. We introduce $Δ$-MOPD, which transfers each teacher's teacher-minus-base logit shift re-anchored at the student's frozen initialization, and compare it with endpoint supervision in both settings while holding teacher selection fixed. We first expose the mechanism that impedes endpoint transfer: inherited base pull can exceed the post-training shift. Removing it reduces the teacher-term norm ratio and target--student KL. Across our experiments, the results suggest that shift targets are particularly useful when teacher signals are combined at a state. With three composed teachers, $Δ$-MOPD exceeds endpoint composition by $4.11$ Math and $1.95$ five-benchmark points; with two, it matches endpoint accuracy. Under phased routing, it achieves higher mean performance in both phase orders and reduces the observed order gap from $10.50$ to $6.42$ points. Under interleaved routing, where each update involves one teacher, the two targets perform comparably. The phased results provide supporting evidence that the benefit may extend to signals accumulated across training phases. Target construction is thus an independent design axis in MOPD, complementary to teacher selection. The experiment repo could be find at https://github.com/HJSang/Delta-MOPD

0.24

Relocating Nonlinearity: How a Downstream Learner Reshapes What Genetic Programming Must Evolve

cs.NE cs.LO ·2026-10-07 ·Nam H. Le
Abstract

Genetic programming was conceived as a way of evolving solutions: the program is the answer, and fitness is the error of its own output. A substantial line of work instead makes the program an input to a separate learner, so fitness measures the learner's output rather than the program's. Which learner to attach matters, with no account of what decides it. What makes a target hard for genetic programming is how much nonlinearity the program must build by composing primitives; whatever nonlinearity the learner supplies, the program need not. The quantity to measure is therefore how much of a target's nonlinearity a learner can take over, though on continuous benchmarks it can only be estimated. Here we show that how much the program must still build is decided by which learner is attached, and is measurable on the programs themselves. Moving to Boolean domains, where a target's nonlinearity is exactly its Fourier degree, we tune that degree from one to six with everything else fixed. Conditioned on success, a linear learner forces the evolved program to the target's degree exactly, at one through six without exception over thirty runs per setting, while tree ensembles succeed with far simpler programs and four times as often. This gives the field two things: a learner can be chosen from a target's structure instead of its reputation for difficulty, and the degree of the evolved program is a diagnostic free to compute during any run. Our control target has lower degree than the parity problems yet gains nothing from a nonlinear learner, while targets reducible to a simple statistic gain a great deal. The latter comes with a warning: evolution internalises what the learner supplies in only four of sixty-three conditions, and grows more dependent on it in forty-four, so the more capable the learner, the less of the model is legible in the program.

0.24

Pre-training of Bayesian Optimization Algorithm through Bayesian Optimization

cs.LG stat.ML ·2026-10-07 ·Satoshi Katayama, Shoyo Hunt, Shintaro Masuda, Masayuki Karasuyama
Abstract

Bayesian optimization (BO) is widely used as a standard approach for expensive black-box optimization. However, BO algorithms often involve parameters that must be specified in advance, and their performance can strongly depend on these choices. We propose a framework for optimizing such parameters using sample paths drawn from a Gaussian process (GP) inferred from the information available at the start of BO. We use cumulative regret as the performance metric for a BO algorithm. By running the BO algorithm on the generated sample paths, we obtain an empirical estimate of its expected cumulative regret for a given parameter configuration. Optimizing this estimate allows us to identify parameter configurations that, given the currently available information, are expected to achieve low cumulative regret. Since this parameter optimization is itself a black-box optimization problem, we employ another BO procedure to solve it, which we refer to as outer BO. Through experiments, we demonstrate that the proposed framework can effectively select parameter configurations that achieve strong performance among a range of candidate configurations.

0.24

Revisiting Explainable AI through Model-Independent Concept Dictionaries

cs.LG stat.ML ·2026-10-07 ·Thomas Schnake, Doreen Schöppenthau, Alexander Meyer, Jacques Corbeil, Klaus-Robert Müller +1 more
Abstract

Modern applications of AI rely on increasingly complex models. Explainable AI (XAI) has emerged as a set of techniques aimed at improving model transparency. However, existing XAI methods typically assume input features to be inherently interpretable, or they rely on intermediate internal abstractions that are difficult to characterize and highly architecture-specific, hindering consistent use across models. To address these limitations, we propose DictXAI, a method that defines concepts directly in the input domain via a dictionary---a large, potentially overcomplete set of predefined elements, each carrying an interpretable meaning. Technically, DictXAI first computes a sparse code of the input and then attributes the model's prediction to the associated dictionary elements. We demonstrate the actionable nature of DictXAI explanations, showing that they can attribute AI malfunctions (e.g., Clever Hans effects) directly to identifiable artifact patterns in the data, while fostering human-AI alignment on intricate biomedical signals. We further demonstrate our method's ability to operate across a wide variety of dictionaries, including learned image bases, analytically defined waveforms for electrocardiography, and experimentally acquired dictionary elements. Overall, our results show that DictXAI provides more interpretable, actionable, and architecture-agnostic insights than classical XAI or existing concept-based approaches.

0.24

NeuralZip: Reusable Setup for Fast Lossless Compression

cs.LG ·2026-10-07 ·Martín Bravo, Samuel Horváth, Gonzalo Navarro, Andrés Abeliuk
Abstract

Lossless compression can reduce the storage and movement of model weights without changing their floating-point values, but repeated statistical analysis and code construction add computational overhead. We study whether the statistical structure of exponents can be prepared once and reused. For this, we introduce NeuralZip, which groups chunks with similar exponent distributions, shares Huffman codes, and selectively represents recurring exponent tuples using packed exponents, thereby achieving additional moderate compression ratios. A setup chooses these representations before subsequent encodings, while every encoding still processes the current tensor values. In floating-point model checkpoints, post-setup compression is 1.81-21.33$\times$ faster than the baselines and achieves exact bit-to-bit reconstruction. We show that this setup can be precomputed and transferred from another compatible architecture, preserving similar compression ratios and avoiding the need to amortize setup costs. Therefore, compression adaptation is transferable and reusable. Training checkpoints demonstrate continued reuse as the weights evolve. Finally, GPU experiments reduce active memory usage by up to 27.5$\%$ while reproducing the logits exactly.

0.24

RECAST: Learning to Compute the Right Context through Adaptive Evidence Routing

cs.AI ·2026-10-07 ·Yilun Hao, Krishna Sayana, Isabella Ye, James S Ren, Sukhdeep Sodhi +2 more
benchmark
Abstract · 35 pages, 2 figures, 13 tables

Large language models are increasingly applied to tasks grounded in long, heterogeneous information sources. Conventional Retrieval-Augmented Generation (RAG) relies on fixed similarity-based retrieval, while agentic variants adapt queries and tool use but remain largely retrieval-centric. However, in many tasks, the evidence required for a solution is not explicitly present in any single source item. Instead, it must be derived through filtering, aggregation, or computation across multiple source items. In this work, we introduce RECAST (Routing Evidence through Computation, Access, and Synthesized Tools), a learned framework that formulates evidence construction as a sequential decision process over heterogeneous retrieval and computation operations, allowing evidence to be actively derived rather than merely retrieved. A lightweight RouterLM iteratively selects and formulates primitive operations or specifies customized operations for a frozen CompilerLM to translate into executable code. Once it judges the evidence sufficient, RouterLM passes the accepted evidence to a frozen AnswerLM to produce the final solution. We train RouterLM with supervised fine-tuning (SFT) followed by group relative policy optimization (GRPO). Across six heterogeneous benchmark families, RECAST achieves a mean success rate of 75.6%, outperforming the strongest large-model baseline by 15.9%. Moreover, training enables the Qwen3.5-9B RouterLM to outperform a training-free Gemini 3.5 Flash RouterLM by 5.0%. On three held-out benchmarks, RECAST improves over the strongest baseline by 15.0% on average, demonstrating strong zero-shot generalization across tasks and heterogeneous source representations.

0.24

How to train your model organism

cs.LG ·2026-10-07 ·Xilin Wang, David Bau, Byron C. Wallace
position
Abstract

Model organisms of alignment-relevant behaviors (e.g., backdoors, sycophancy, spurious correlations) have emerged as a key tool for evaluating whitebox interpretability techniques. We argue that the prevailing practice of training model organisms to a single objective of installing the target behavior is insufficient and propose validating model organisms with respect to three objectives with associated metrics: target-behavior installation, general-capability preservation (i.e., parametric knowledge, chat quality), and output naturalness (i.e., CoT and activations). We re-visit two publicly released organism suites using this validation framework and show that (1) chat quality and CoT naturalness degrade substantially across training recipes, and (2) validation metrics predict how well interpretability methods recover the installed behavior, e.g., a logit lens readout covaries with an organism's general capabilities. We introduce a multi-objective training approach based on model merging to train more realistic model organisms. Finally, on a new suite of model organisms targeting demographic biases in clinical reasoning, we compare training recipes and find that DPO training stays closer to the base model than supervised finetuning, and the proposed model optimization approach better preserves capabilities and naturalness. Auditing this suite with an investigator agent, we again observe validation metrics tracking bias recovery. In sum, training methods shape the interpretability conclusions an organism supports, and we argue that one should consider multiple objectives to draw generalizable conclusions about interpretability methods using (realistic) model organisms.

0.24

KGATE : a Knowledge Graph Embedding Training Environment

cs.LG cs.AI ·2026-10-07 ·Benjamin Loire, Galadriel Brière, Célia Brahimi, Antoine Toffano, Anaïs Baudot
Abstract · Main paper (7 pages, 1 figure) and supplementary materials (4 pages, 1 figure, 3 tables) provided

Knowledge graph embedding (KGE) models encode the entities and relations of a knowledge graph into a low-dimensional latent space, enabling tasks such as classification or link prediction. Most KGE models follow an autoencoder architecture, in which an encoder projects the knowledge graph into the latent space and a decoder reconstruct it. Combining both encoder and decoder components is increasingly needed, yet existing libraries rarely support complete autoencoders, are often unmaintained, rely on undocumented default hyperparameters, and produce results that cannot be compared across libraries. Here we present KGATE (Knowledge Graph Autoencoder Training Environment), a modular Python library built on PyTorch Geometric and TorchKGE. KGATE lets users assemble initializers, encoders, decoders, losses, negative samplers, and evaluation metrics as building blocks, or plug in their own block. KGATE includes a preprocessing procedure that controls data leakage, a builtin training pipeline, and reproducibility by design. Benchmarks against six existing KGE libraries show that KGATE training time is comparable with the fastest libraries while offering a broader set of features.

0.24

Broadly Applicable Approximate MCMC for Switching Stochastic Differential Equations Using Uniformization and Time-Conditioned Factorized Neural Likelihood Estimation

cs.LG stat.ML ·2026-10-07 ·Shion Hosoda, Michiaki Hamada
Abstract

Switching stochastic differential equations (SSDEs) describe continuous-time dynamics whose parameters switch according to a latent regime process that follows a continuous-time Markov chain (CTMC). By allowing dynamics to change between regimes, SSDEs represent heterogeneous system behavior and have been applied across diverse fields. However, Bayesian inference for SSDEs remains difficult, and existing SSDE inference methods have limited applicability, with restrictions such as noise-free observations, univariate states, linear drift, or state-independent diffusion. In this study, we propose an approximate Markov chain Monte Carlo sampler for SSDEs using uniformization and factorized neural likelihood estimation (FNLE), a simulation-based inference method. Uniformization provides an exact representation of the CTMC but requires SDE transition densities over arbitrary time intervals. We approximate these densities by training a time-conditioned FNLE model. The resulting sampler is broadly applicable to SSDEs without requiring analytically tractable transition densities. In synthetic-data experiments, our method recovered regime paths and parameters for three SSDE models for which previous methods have limited applicability. We also applied our method to a real dataset and detected a regime transition.

0.24

EntroPrefill: Renyi-Guided Context Pruning with Conditional Stability Guarantees for Retrieval-Augmented Generation

cs.LG stat.ML ·2026-10-07 ·Inbasekaran S
theory
Abstract · 16 pages, 0 figures. Theoretical framework and conditional stability guarantees for context pruning; no empirical evaluation or benchmarks a…

Mid-prefill pruning can reduce the sequence processed by deeper transformer layers, but attention concentration alone does not certify that discarded context is dispensable. We formulate EntroPrefill as a Renyi-guided proposal mechanism coupled to explicit constraints on discarded attention mass. Sink-isolated, regularized head pooling respects grouped-query attention while exposing a quantitative trade-off between specialization and worst-head coverage. We derive a mixture-to-head deletion envelope, a computable upper bound on feasible token removal, and a finite-sample observer guarantee that remains valid when the pruning layer is selected adaptively. We then establish a conditional transformer perturbation bound with explicit sufficient Lipschitz constants and a first-token decision-margin corollary. A counterexample shows why shallow observations alone cannot imply an unconditional future-output guarantee. The systems analysis distinguishes query-head unions, physical page allocation, and KV-transfer payload, and gives an arithmetic break-even condition for pruning. This manuscript is theoretical in scope: it defines the procedure, its assumptions, and its formal limits, but does not report measured acceleration or task-accuracy preservation. Experiments are reserved for subsequent validation of the assumptions, approximation tightness, and end-to-end resource trade-offs.

0.24

Learning Situation-Conditioned Thinking Policies for Long-Term LLM Agents

cs.AI cs.RO ·2026-10-07 ·Hong Su
Abstract

Long-running autonomous agents must reuse accumulated reasoning experience without allowing explicit historical memory and LLM context to grow indefinitely. However, existing memory mechanisms mainly retrieve, summarize, or compress past content and do not directly learn when particular kinds of thinking should be activated or discover new thinking knowledge from temporally dispersed experiences. This paper proposes a situation-conditioned thinking memory framework that transforms historical reasoning experience into a lightweight policy for predicting what should be thought about in the current situation, while leaving detailed reasoning to a large language model. Situations may represent temporal or spatiotemporal evolution rather than only current states. Temporary experiences are also periodically analyzed across multiple independent episodes to identify repeated long-range regularities, which are consolidated into new thinking knowledge and further internalized by the lightweight policy. Experiments show that the learned policy achieves 1.000 F1 on temporal-rule generalization, improves DeepSeek reasoning F1 from 0.789 to 0.868, reduces online processing time from 0.3636 ms to 0.0382 ms per query at 30,000 historical situations, and reaches 1.000 relation-discovery F1 and future-thinking accuracy after sufficient repeated cross-experience evidence.

0.23

Multi-Bandwidth Distribution Matching Distillation: On the Equivalence of Distribution Matching Distillation and Drifting Models

cs.LG cs.CV ·2026-10-07 ·Jialin Zhu, Xing Liu, Feixiang He, He Wang
theory
Abstract

Researchers are exploring effective one-step generative model continuously, and, Drifting Models (Deng et al., 2026), demonstrate great potential in one-step generation recently. There are works that reveal the connection between Diffusion & Flow Style Generative Models (DFSGMs) (Ho et al., 2020; Song et al., 2020a;b; Lipman et al., 2022; Liu et al., 2022) and Drifting Models (Li & Zhu, 2026; Lai et al., 2026; Turan et al., 2026). But no one has yet established a precise correspondence between the Drifting Model and the widely used distillation method- Distribution Matching Distillation (DMD/DMD2) (Yin et al., 2024b;a) to the best of our knowledge, even though their optimization objective formulas are virtually identical. In this paper, we prove that by converting the velocity-field / noise-field from the pre-trained DFSGMs into the attraction force field in Drifting Models and estimating the repulsion force field from the generative distribution, training the Drifting Model is naturally equivalent to the Distribution Matching Distillation. With this equivalent concept, we propose an improved method based on DMD from the Drifting Model's perspective- Multi-Bandwidth Distribution Matching Distillation (MBDMD).

0.23

Boundary-aware Reinforcement Learning for Hypercube State Spaces via Deterministic Policy Gradient

cs.LG math.OC ·2026-10-07 ·Lijun Bo, Yijie Huang, Chenhao Lu
theory
Abstract

We develop a continuous-time deterministic policy gradient framework for reinforcement learning with reflected state dynamics, where the state process is governed by a controlled reflected stochastic differential equation on a hypercube. Under suitable regularity assumptions, we establish the connection between the value function and the Neumann Bellman equation, introduce an advantage-rate function that yields a deterministic policy gradient formula, and prove the martingale characterization theorem. Motivated by these theoretical results, we propose a continuous-time deep deterministic policy gradient algorithm for reflected stochastic systems, in which the Neumann boundary condition is imposed via either soft penalization or hard architectural constraint. We further quantify the discrepancy between the ideal continuous-time dynamics and the discretely sampled exploratory dynamics executed in practice, showing that the error decays as the time grid is refined and exploration noise vanishes. Our experiments on reservoir control problems illustrate the effectiveness of the RL framework, highlighting that boundary-aware methods substantially reduce Neumann boundary residuals and enhance learning stability.

0.23

TR-PTQ: High-Accuracy Integer-Only Transformer Post Training Quantization via Taylor Region Reformulation

cs.LG ·2026-10-07 ·Eliyahu Levy, Adam Teman, Yoni Pugachov
Abstract

Post-training quantization (PTQ) enables efficient deployment, yet transformer architectures remain challenging to quantize due to nonlinear layers. While existing methods attribute accuracy loss to insufficient numerical precision, often necessitating floating-point fallbacks, we demonstrate that degradation is actually driven by specific structural error sources. We find that learned scale parameters in normalization layers and compounded approximations in GELU are the primary error contributors, whereas SoftMax remains inherently robust to aggressive quantization. To address these bottlenecks, we introduce TR-PTQ, a unified integer-only formulation using shared Taylor Region (TR) exponential and logarithm primitives. This approach allows computationally expensive operations, including division and square roots, to be performed entirely in the log-domain via standard integer arithmetic. Combined with a calibration-free, outlier-aware optimization for LayerNorm parameters, our method eliminates the need for floating-point hardware units for nonlinearities, achieving less than 1.5\% absolute accuracy degradation across vision and language benchmarks.

0.23

Whose Ground Truth? Embracing Ambiguity in Human-Centered AI

cs.AI cs.LG ·2026-10-07 ·Jingyao Wu, Mohammad Tariqul Islam, Per Rådberg Nagbøl, Julie Gerlings, Giovanni Leoni +3 more
publishedposition
Abstract · Accepted to NeurIPS 2026 Trustworthy AI for Good Workshop

As AI systems increasingly interact with people and make decisions about them, understanding human interpretations becomes an important part of developing human-centered AI. Conventional machine learning and AI systems are largely developed under the assumption that a single definitive ground truth exists, with variability in human annotations often resolved through aggregation or treated as noise. However, for many human-centered tasks, human interpretation is inherently ambiguous, and multiple interpretations of the same input may be simultaneously reasonable and valid. Reducing such ambiguity to a single target risks overlooking meaningful information about the diversity of human perception, judgment, and experience. In this position paper, we call for a shift towards modeling the interpretation space of plausible human judgments, while distinguishing meaningful ambiguity from annotation noise. We argue that this perspective should guide how AI systems are represented, learned, evaluated, deployed, and governed, supporting more human-centered AI that better reflects the diversity of human interpretation.

0.23

Expected Sample Complexity in Multi-Armed Bandits

cs.LG stat.ML ·2026-10-07 ·Nadav Sukenik, Nadav Merlis
hot direction
Abstract

Sample complexity is a widely used metric in sequential decision-making problems, defined as the number of suboptimal decisions during the interaction between the agent and an environment. We study the sample complexity of stochastic multi-armed bandit problems and introduce the expected sample complexity performance measure, analyzing it in a novel framework called approximately correct in expectation (ACE). We show that ACE guarantees imply almost sure convergence to the optimal expected reward, in contrast to high-probability guarantees found in other frameworks, and also show how to convert ACE guarantees into explicit expected regret bounds. We further show that, in contrast to existing measures, deterministic algorithms cannot obtain favorable ACE bounds, and analyze stochastic algorithms in two settings: when the allowed suboptimality level $ε$ is known to the algorithm and when it is unknown. In the former, we devise an explore-then-$ε$-greedy algorithm, and in the latter, we analyze the expected sample complexity of Thompson sampling. Finally, we establish nearly matching lower bounds for both settings, showing that the algorithms are tight in $ε$ and proving a performance separation between the two regimes.

0.23

Can the Library Speak? Self-Advertised Method Selection for Formal Proving

cs.AI ·2026-10-07 ·Xiaopeng Yuan, Suijin Wang, Yanli Wang, Haibo Jin, Peng Kuang +3 more
theory
Abstract · 17 pages. Preprint

LLM-based formal provers can retrieve relevant lemmas and prior proofs, but relevance alone does not say whether a mathematical method can be used on the current theorem. A method has prerequisites, a target, an intended action, and obligations that its use leaves to prove. Methods that look equally related to a theorem may therefore differ substantially in whether they offer a plausible next step. We formulate this as an applicability-aware method-selection problem and introduce self-advertisement: before candidates are ranked, a model generates a problem-specific proposal for each one, stating what part of the goal it targets, what action it would take, and what conditions that action requires. We organize 82 reusable methods from Putnam 2000-2014 as Method Contracts, which pair applicability descriptions with Mathlib anchors, a checked example or scaffold, and expected proof obligations. A single batched call elicits proposals across the library; vague or unsupported proposals are demoted, yielding a ranked shortlist accompanied by inspectable claims about each candidate's use. We analyze when similarity-based representations cannot distinguish methods with different applicability, how errors in applicability estimates affect shortlist quality, and what a checked scaffold guarantees under its stated assumptions. Against lexical, embedding, and embedding-plus-LLM reranking baselines, self-advertisement achieves 95.0% hit@5 on Putnam 2015-2025, compared with 84.2% for the strongest reranker. On IMO ProofBench, it achieves 91.7% compared with 88.3%. These results indicate improved coverage of annotated methods in the retrieved shortlists, particularly on Putnam.

0.23

PairAudit: Guiding Human Review with Graph Tokens under Distribution Shift

cs.LG cs.CR ·2026-10-07 ·Jiran Tao, Binyan Jiang
Abstract · 22 pages, 3 figures

Intrusion detectors can confidently misclassify attacks that were not seen during training. Human review can correct these errors, but only a limited number of cases can be checked. Uncertainty-based review may overlook confident errors, while anomaly scores alone do not show whether changing the review plan will correct more errors. We introduce PairAudit to find overlooked errors and improve review under a fixed budget. Its graph tokens capture prediction patterns across connected nodes. Rather than building another predictor through feature aggregation, PairAudit uses unusual relational patterns to uncover potential errors in existing predictions. Human feedback then helps decide whether these findings justify changing review priorities. Experiments across security tasks show that PairAudit corrects more errors on average than uncertainty-based review, including more errors on unseen attacks. These gains account for all review costs and do not require retraining the detector.

0.23

Lightweight and Versatile Learned Optimization by Recombination of Gradient History

cs.LG ·2026-10-07 ·Minyoung Choi, Dalta Imam Maulana, Wanyeong Jung
Abstract

This paper presents a lightweight and versatile learned optimizer that dynamically recombines gradient history, represented as averages over disjoint time spans. The optimizer reduces the prediction space to one scalar coefficient per gradient average, shared by multiple parameters. Progressively averaging older gradients minimizes memory cost of long history, while keeping their contributions independently accessible. A 37k-parameter network trained in 0.87 GPU-hours generalizes zero-shot to unseen tasks, lowering validation loss by 9.1% and 0.4% on BERT-Tiny and GPT-Tiny, and improving test accuracy over Adam by 3.5 %p on a Vision Transformer and by 2.7 %p on average across nine graph models, with FLOPs overhead as low as 0.3%.

0.23

SkillSandbox: Skill Verification via Dynamic Scenario Synthesis

cs.AI cs.CL ·2026-10-07 ·Serin Kim, Kwangwook Seo, Dokyung Song, Jinyoung Yeo, Dongha Lee
Abstract

Self-evolving agents distill task-solving experience into skills for future reuse, but these skills can encode incorrect procedures or non-transferable knowledge. It is therefore critical to verify each skill's reusability: whether its guidance remains useful beyond the experience from which it was distilled. Such verification requires observing how a skill affects execution in new tasks, yet existing tasks may not expose the situations where the target skill can actually be exercised. To construct such situations, we propose SkillSandbox, a framework that dynamically synthesizes a task and its environment for each skill that are skill-relevant yet novel. A Proposer specifies the conditions to preserve and the source-specific details to vary, a Builder constructs an executable scenario, and a Verifier compares executions with and without the skill. The Verifier assesses executability, utility, and efficiency to assign a Keep or Reject verdict, determining whether the skill enters the library. Across ALFWorld and WebShop with three models, SkillSandbox consistently yields the strongest downstream performance and improved execution efficiency. Further analyses examine whether these gains reflect accurate assessment of skill reusability and identify which components of SkillSandbox contribute to them.

0.23

Structure alone supports efficient visual computation in the Drosophila visual system

cs.LG q-bio.NCcond-mat.dis-nn ·2026-10-07 ·Eudald Correig-Fraga, Roger Guimerà, Marta Sales-Pardo
Abstract

Understanding the extent to which measured synaptic wiring determines computation remains a central challenge. Here, we couple the proofread adult Drosophila melanogaster connectome to an anatomically faithful model of its eye. Visual information is inputted in the eye model, then passed to the connectome, and finally read from a Kenyon-cell-centered linear decoder. This creates a connectome-only model in which the anatomical graph and eye geometry are fixed and only scalar synaptic gains and neuronal thresholds may be learned. The model supports multitask vision, including color discrimination, shape classification, and numerical discrimination that follows a ratio-dependent scaling characteristic of approximate number perception. To test whether precise connectivity is consequential under wiring economy, we compare the biological graph to randomized ensembles that increasingly preserve biological synaptic constraints. At matched wiring cost, the biological network consistently yields higher accuracy, whereas less constrained rewiring surpasses it at the cost of inflated wiring. These findings indicate that the measured connectivity and eye geometry jointly set efficient operating points for visual computation.

0.23

The surprising simplicity of forecasting quantum chaotic many-body dynamics

cs.LG quant-phcond-mat.stat-mech ·2026-10-07 ·Francesco Perciavalle, Agostino Gallo, Francesco Plastina, Gianluigi Greco, Nicola Lo Gullo +1 more
benchmark
Abstract · Francesco Perciavalle and Agostino Gallo contributed equally to this work. 11 pages, 5 figures

Accurately simulating chaotic quantum many-body dynamics remains a major computational challenge for classical methods, due to the rapid buildup and spatial spreading of entanglement during the evolution. This raises the question of whether machine learning can provide an effective alternative for predicting quantum dynamics. We address this question by formulating quantum dynamics as a time-series forecasting problem, using a few-qubit PXP chain, realizable with Rydberg-atom arrays, as a benchmark. By varying the initial state, the system spans dynamical regimes ranging from ergodic behavior to quantum many-body scarring, providing a controlled setting for testing forecasting models across qualitatively different dynamics. We compare two contrasting architectures: an expressive nonlinear Transformer and DLinear, a simple linear forecasting model. The Transformer accurately predicts dynamics in the more ergodic regime, but its performance progressively deteriorates as the initial state approaches the scarred limit. In contrast, DLinear remains accurate across the entire family of initial states, with its main deviations consisting of small high-frequency oscillations that have little effect on the overall prediction error. Remarkably, these results show that observables generated by complex quantum many-body dynamics can be forecast with high accuracy through a simple linear mapping from past to future observations. This reveals that the complexity of the underlying quantum evolution need not translate into an equally complex forecasting problem.

0.23

Transferability of Learned States in Neural PDE Solvers

cs.LG ·2026-10-07 ·Shunye Wang, Haochen Wen, Shuo Li Liu, Xuanyi Wang, Lihao Liu +1 more
Abstract

Assessing useful reuse in neural PDE solvers is challenging: final accuracy can reflect source learning and target-time computation. Our reuse contract separates solution accuracy, learning contribution, and numerical utility through paired state comparisons, matched target information and budgets, and cost accounting. A literature audit extracts 18 version-specific protocol records from 12 papers, documenting retained states, target-time resources, and reported controls. For a fixed linear system and residual tolerance, we construct two initial guesses with identical solution-error, energy-error, and residual norms, reaching the same solution with different conjugate-gradient (CG) iteration counts. Across 240 source-training trajectories, two linear PDE families, Fourier neural operators and convolutional networks, a fixed predictor's benefit reverses across correction algorithms. Among pairs with both relative prediction errors less than or equal to 5 percent on 64 in-distribution tasks (63 by 63 interior grids), reductions in all three norms accompany more CG iterations, at mean taskwise rates of 23.5 percent and 23.9 percent in two libraries. Work-based selection saves 2.50-3.33 CG iterations on held-out in-distribution tasks; matched adaptation demonstrates finite-budget pretraining value. Independent batches confirm a 0.73 percent complete online saving for one physics-trained Fourier neural operator against zero-initialized Poisson-preconditioned CG. Reuse requires matched state comparisons and downstream computational evidence.

0.23

CERO: Where and When to Allocate Rollouts for RL Post-Training

cs.LG cs.AI ·2026-10-07 ·Yiming Zong, Yige Wang, Xing Hu, Jiashuo Jiang, Zuo-Jun Max Shen
Abstract

Adaptive rollout methods for group-relative reinforcement learning typically allocate a fixed per-update budget across prompts. We instead study how to coordinate a finite rollout budget over the entire training horizon. We formulate this problem using a concave surrogate utility of cumulative prompt exposure and introduce CERO, an online primal dual scheduler for prompt admission and budget pacing. In our experiments, each admitted prompt receives a fixed-size response group. CERO instead adapts which prompts are selected, how often they are revisited across rounds, and how many groups are generated in each round. A compact Fenchel representation linearizes the dependence on cumulative exposure, while projected online gradient descent updates prompt-specific supporting slopes and a shared budget price using reward-variation feedback and budget deviations. We establish pathwise guarantees for the surrogate allocation objective against fixed-rate and same-path time-varying benchmarks, with explicit terms for proxy discrepancy and rate variation. Under matched training-response budgets, CERO attains the highest avg@16 macro-average on each of three backbones across five mathematical reasoning benchmarks. Mechanistic analyses link CERO's prompt choices to within-group reward contrast, while multi-seed ablations show gains from adaptive pacing over both uniform and preset spending schedules.

0.23

What can linear attention learn from nonlinear teachers in-context?

cs.LG stat.ML ·2026-10-07 ·Mary Letey, Arman Rysmakhanov, Yue M. Lu, Cengiz Pehlevan, Jacob Zavatone-Veth
Abstract

Linear attention is a tractable model for understanding the mechanisms governing in-context learning in transformers. For linear regression tasks, recent asymptotic analyses have characterised its learning and generalisation behaviour. We extend this theory to nonlinear single-index targets, $y=f(x^\top w)+\varepsilon $. Our main result establishes a nonlinearity-noise equivalence: linear attention extracts only the linear Hermite component of $f$, while the remaining nonlinear structure contributes to the generalisation error as effective noise. This reduction allows results from the corresponding linear theory to be transferred to nonlinear tasks. We illustrate its implications for finite pretraining data and for the transition from task memorisation to task generalisation as task diversity increases. These results identify a limitation of the reduced linear-attention model and provide a tractable starting point for studying nonlinear in-context learning.

0.23

DUDA-Bench: Benchmarking LLM Agents on Multimodal Data-Driven Urban Diagnosis

cs.AI ·2026-10-07 ·Yizhi Song, Hang Ni, Weijia Zhang, Hao Liu
benchmark
Abstract

Urban diagnosis integrates heterogeneous observations to identify urban problems, localize affected areas, and investigate contributing factors, informing evidence-based urban planning and management. However, its reliance on labor-intensive, case-specific expert workflows limits scalability and reuse, motivating the exploration of agent-based execution. To evaluate this capability, we introduce DUDA-Bench, a hierarchical and interactive benchmark that formalizes data-driven urban diagnosis as a multi-stage agent workflow. It comprises 86 atomic and 22 workflow tasks spanning four analytical stages, grounded in multimodal data from 12 cities covering five urban problem types. Evaluations of seven backbone models and five agent systems reveal a substantial gap between isolated analytical competence and end-to-end diagnosis, with system benefits varying across backbones. Trajectory analysis shows that unresolved evidence gaps propagate across stages, while successful recovery involves revising assumptions and actions using feedback. These findings highlight limitations in coordinating analytical capabilities across stages, particularly adaptive planning, evidence integration, and verification. More broadly, DUDA-Bench provides a framework for translating expert analytical workflows into hierarchical agent tasks and process-aware evaluation, supporting systematic assessment of end-to-end analytical capabilities.

0.23

Continual Graph Multi-Agent Reinforcement Learning

cs.LG ·2026-10-07 ·Tommaso Marzi, Ahmed Hendawy, Jan Peters, Carlo D'Eramo, Andrea Cini +1 more
benchmark
Abstract

In Continual Multi-Agent Reinforcement Learning (CMARL), agents learn cooperative policies across sequences of tasks, aiming to adapt effectively to new tasks while preserving the ability to solve previously encountered ones. In many applications, tasks differ in their underlying structure, which can represent, for example, distinct operational conditions or target configurations (e.g., different network topologies in power grids or arrangements in formation control). Existing CMARL methods lack dedicated mechanisms to leverage this structural information when learning new tasks, failing to promote transfer and mitigate forgetting. To fill this gap, we propose Continual Graph Multi-Agent Reinforcement Learning (CGMARL), a novel framework for CMARL problems in which task sequences are mapped into a series of attributed graphs, each modeling a task-specific structure. In CGMARL, each graph determines the environment dynamics (next states and/or rewards) and the number of agents for the corresponding task. Then, we present Graph-based Formation (GRAFO), the first CGMARL benchmark, and show how forgetting arises in this setting. Finally, to address this limitation, we propose Frozen Graph Encoder (FROG), a method that relies on a frozen graph backbone to preserve past structural information in graph-based CMARL policies. Experiments on GRAFO show that pairing FROG with existing CL methods substantially improves performance on multiple CGMARL scenarios.

0.22

Fluctuations of Nonlinear Observables in Mean Field Neural Network Training

cs.LG math.PRmath.STstat.ML ·2026-10-07 ·Arnaud Descours, Geoffrey Lacour
theory
Abstract

Mean field limits describe the training dynamics of wide neural networks through the evolution of the empirical distribution of their parameters. Although functional central limit theorems characterize the asymptotic fluctuations of this distribution, quantities of practical interest are typically nonlinear observables of the parameter distribution rather than the distribution itself. In this work, we show how these mean field fluctuations propagate to finite dimensional nonlinear observables for shallow neural networks trained by stochastic gradient descent. Working in the weighted Sobolev space in which the limiting fluctuation process is constructed, we apply a functional Delta method under ordinary Fr{é}chet differentiability, without requiring Lions derivatives with respect to the measure variable. We obtain a central limit theorem for the observables and, under a suitable representation of their differentials, an explicit covariance formula inherited from the underlying mean field fluctuation theory. We also study whether prescribed quantities of interest can be recovered from the selected observations. Under a constant rank assumption, we prove that a quantity of interest factors locally through the observation functional if and only if, throughout a neighborhood, the kernel of the differential of the observation is contained in that of the quantity of interest. Thus, a differential condition expressed directly in the ambient Sobolev space yields an exact nonlinear local factorization. These results provide a framework both for quantifying finite-width uncertainty on observable, statistically or physically meaningful quantities and for assessing whether the chosen observations contain the information required to identify them.

0.22

Thinking in Depth: Retrospective Inference for Tabular Foundation Models

cs.LG ·2026-10-07 ·Hao-Run Cai, Si-Yang Liu, Zi-Jian Cheng, Kun-Yang Yu, Jin-Hao Sheng +6 more
Abstract

Tabular foundation models (TFMs) are pretrained across diverse tabular tasks and make predictions on a new table at inference time using its labeled examples as context. Most recent TFMs perform such in-context prediction with stacked Transformer layers, repeatedly transforming how examples are represented and compared. By tracing individual queries through several strong TFMs, we find that predictive refinement is highly uneven across depth and is often concentrated in later layers. This uneven refinement motivates us to reconsider how intermediate representations are constructed and reused throughout the network. We introduce Retro, a tabular foundation model based on retrospective inference, where later stages can explicitly revisit and recombine intermediate information produced earlier in the network. Retro organizes this process around two complementary operations: which intermediate information to revisit, and how the resulting contextual update should be shaped for each query. Attention Residuals address the former by adaptively reweighting contributions from different depths, while query-conditioned Gated Attention addresses the latter by modulating the attention output element-wise across representation dimensions. Our analysis shows that Retro shifts predictive refinement earlier and more broadly across depth, with different stages revising different subsets of queries in a pattern suggestive of multi-view refinement. Across TabArena, TALENT, and RelArena, Retro ranks among the top three and lies on the Pareto frontier. These results indicate that directly reusing intermediate representations provides a practical way to better exploit depth in TFMs.

0.22

Calibrating Ambiguity Set via Diagnostic Transport for Distributionally Robust Optimization

cs.LG stat.MLmath.OC ·2026-10-07 ·Wenbin Zhou, Elizabeth Cucuzzella, Shixiang Zhu
Abstract

Distributionally robust optimization (DRO) protects decisions against distributional uncertainty by optimizing over an ambiguity set, but poorly aligned set geometry can require large radii and yield overly conservative decisions. We introduce diagnostic-transport DRO (DT-DRO), which uses held-out calibration data to adapt the ambiguity-set geometry to observed predictive errors. DT-DRO uses the conditional probability integral transform cumulative distribution function to diagnose systematic probability misallocation and translates this information into an outcome-level transport that jointly adjusts the ambiguity-set center and ground cost. The resulting formulation admits a computationally tractable dual reformulation. Theoretically, we derive valid ambiguity radii and decision-risk guarantees that tighten as estimation and approximation errors vanish, and show that DT-DRO can eliminate the nonvanishing robustness floor caused by model misspecification. Synthetic experiments and a power-outage application demonstrate improved decision quality, particularly under structural and tail misspecification.

0.22

LiveMACE: Process-Aware Evaluation of LLM Agent Capabilities in Evolving Markets

cs.AI cs.CL ·2026-10-07 ·Jun Zhao, Leiming Fu, Yanbo Wen, Yiding Wang, Xuantong Liu +7 more
benchmarkcommunity pick
Abstract

Evaluating agents by outcomes alone can obscure the capabilities that produce them. This problem is especially pronounced in evolving environments, where outcomes reflect a closed-loop interaction between agent behavior and changing external conditions. We introduce LiveMACEBench, a process-aware benchmark that uses live financial markets as a naturally evolving testbed for persistent LLM agents. Five frontier LLMs operate along continuous trajectories under matched Tool Use, Persistent Memory, Rule Following, and Multi-Agent Collaboration configurations. We evaluate them through both realized outcomes and mechanism-specific diagnostics derived from complete decision traces. Across 30 days of live evaluation, we find a pronounced outcome-capability gap: realized returns often diverge from capability-specific measurements, and similar outcomes can arise from markedly different patterns of mechanism use. Trace-level diagnostics further expose distinct bottlenecks across capabilities, demonstrating that mechanism access, effective mechanism use, and downstream performance are not interchangeable measures of agent capability. LiveMACEBench makes this distinction measurable, turning live markets from a performance leaderboard into a diagnostic environment for agent capability

0.22

An AI-assisted conditioning and geological interpretation workflow for usage in implicit geological modeling

cs.AI physics.geo-phcs.LG ·2026-10-07 ·Stefan Carpentier, Jan Diederik van Wees, Eva de Boever, Jan Niederau, Camille Chapeland +3 more
Abstract · 17 page, 19 figures

Implicit modeling and Relative Geologic Time are geological modeling techniques that enable more efficient, faster, less biased and more reproducible modeling results. For optimal operation, these techniques require many well-constrained input data. In the framework of the Horizon Europe GO-Forward and MOOI WarmingUP GOO projects and to accelerate Implicit modeling, Machine Learning (ML) methods have been tested and implemented in a toolkit for the interpretation of (onshore) seismic data from the shallow to deep range (+- 300 - 3500 m). The goal is to rapidly characterise this depth domain by efficient interpretation of horizons and faults in seismic data. The first step is to improve the signal by applying AI techniques like self-supervised and semi-supervised contrastive learning CNN's for noise reduction and interpolation. Next, horizons and faults are interpreted with minimal use of human-generated training data by using (semi-) self-supervised methods. The resulting developed toolkit supports the application of the implemented algorithms in an efficient workflow. As a first demonstration, the top of the Dutch Maassluis Formation has been interpreted in the Leeuwarden and Waalwijk 3D seismic cubes. Overall, this study demonstrates that AI-assisted interpretation workflows have reached a level of maturity that allows their integration into applied geological modeling and decision-making.

0.22

From Expected Harmfulness to Likelihood: A Probabilistic Reformulation of Jailbreaking LLM Agents

cs.AI ·2026-10-07 ·Juanyang Xu, Zheng Wang, Xingyu Zhao, Siddartha Khastgir, Andi Zhang
Abstract

When the harmfulness of an LLM agent's output can be quantified, a natural jailbreaking objective is to maximize expected harmfulness over admissible input modifications. An alternative approach constructs or selects harmful target outputs and modifies the input to increase their likelihood. We establish a precise connection between these two approaches through a probabilistic reformulation. Specifically, we show that the gradient of the logarithm of expected harmfulness with respect to the input equals the expected input gradient of the model's log-likelihood under a harmfulness reweighted output distribution. This identity provides a unified interpretation of expected harmfulness and target likelihood optimization. Building on this connection, we propose OPUR, a sampling distribution designed to generate highly harmful target outputs and use the resulting samples to guide likelihood-based input optimization. Experiments demonstrate the effectiveness of the resulting method in jailbreaking LLM agents.

0.22

Language Models for Page-Level Layout Decisions in E-commerce Search

cs.LG cs.CLcs.IR ·2026-10-07 ·Varun Joshi, Eva C. Song, ChengXiang Zhai
published
Abstract · Accepted at the OARS Workshop, ACM RecSys 2026

E-commerce search pages are critical touchpoints for millions of online shoppers. While traditional search engines return a ranked list of results, modern E-commerce search pages increasingly incorporate recommender system modules -- for example, secondary stacks that surface alternative product groupings at specific positions. When introduced appropriately, secondary stacks can improve user engagement; however, suboptimal placement may disrupt browsing flow and degrade the primary results. Unlike traditional search ranking, where evaluation techniques such as interleaving are well established, evaluating page-level layout changes e.g., when and where to insert a secondary stack remains challenging without costly online A/B testing. To address this, we study offline methods for evaluating whether a given layout decision -- specifically, the inclusion of a secondary stack at a particular position -- is beneficial to users. We investigate language models as scalable evaluators by comparing direct prompt-based, prompt-derived feature, and representation-based methods. Our results show that representation-based approaches consistently outperform prompt-based judging in predicting user engagement, suggesting they provide a reliable foundation for offline layout evaluation in E-commerce search.

0.22

m-Set Adversarial Bandits with Winner Feedback

cs.LG stat.ML ·2026-10-07 ·Nicolò Cesa-Bianchi, Matteo Papini
hot direction
Abstract

We show upper and lower bounds on the regret of $m$-set adversarial bandits for different utilities (winner reward or sum of rewards) and feedback models (winner index, winner reward, sum of rewards, and their combinations). By comparing to standard bounds for combinatorial and MNL bandits, our results reveal how subtle changes in the setting can have a dramatic impact on the learning rates. Our main technical contributions are the information-theoretic lower bounds on the regret. Experiments on synthetic data confirm our theoretical analyses.

0.22

Automatically Building and Updating a Knowledge Graph of MLIP Models

cs.AI ·2026-10-07 ·Alexis Beer, Liudmyla Klochko, Mathieu d'Aquin
benchmark
Abstract

Complementing the many efforts in providing semantic representations of concepts, notions, and entities in materials science, we report and illustrate a process by which we can automatically build a knowledge graph of the fast evolving field of machine learning applied to the prediction of material properties, focusing on MLIP (Machine Learning Interatomic Potential). This LLM-based process relies on multiple steps, from information extraction in documents and articles to a validation loop using SHACL constraints to detect and correct errors. It is carried out on a model-by-model basis, focusing on the consistency of representation, therefore enabling an iterative construction where the addition of new models is facilitated. We illustrate the process by showing a few interesting aspects that can be queried from a knowledge graph built from the models listed in the Matbench Discovery leaderboard.

0.22

Curating Always-Loaded Context for LLM Agents: A Capacitated Assortment Model with Censored Feedback

cs.AI cs.LGmath.OC ·2026-10-07 ·Zexuan Liu, Yuning Yang, Tiancheng Zhao
theory
Abstract

At the start of every session, LLM agents load a fixed context file, such as $\texttt{AGENTS.md}$. Each loaded token in the file is charged again in every later round of the session, and these files can degrade performance as they grow in size. However, in practice, human or automated curators usually grow these files by appending. We formulate context curation as a capacitated assortment problem. Instructions consume tokens under a finite attention capacity; adding an instruction never raises the compliance of the others, while retained instructions incur a per-session setup cost. We prove an upper bound on the optimal file size, regardless of the number of available candidate instructions, and that appending every instruction with positive standalone value can be arbitrarily worse in net value than selecting an optimal subset. A token budget also limits the loss when the token price is underestimated. We then examine what can be learned from past sessions and how this information can guide decisions to add or remove instructions. Feedback is inherently censored: the benefits and harms of loaded instructions are observable, whereas missing instructions generate feedback only when their absence causes harm. In this setting, we show that deleting instructions ignored by agents can inevitably remove helpful ones. We characterize how much evidence should be collected before adding an instruction. Besides, we bound regret when human reviewers can inspect only a limited number of edits per period. Empirical experiments further show that irrelevant rules drawn from real context files reduce language-model compliance.

0.22

Identifiability of a dissipative knowledge-dynamics model: exact recovery under designed excitation, degeneration on observational data

cs.LG math.OCstat.ML ·2026-10-07 ·Arman Kostanian, Armen Beklaryan
theory
Abstract · Corrected the implicit integration scheme and Proposition 4; revised Lemma 1; specified the practice input, difficulty reference, rank and t…

Human learning is a dissipative dynamical process: mastery accumulates through practice, decays through forgetting, and propagates across interdependent concepts. We model it as a nonlinear dissipative system of ordinary differential equations whose parameters are mechanistically meaningful (a transfer-of-practice matrix, per-concept forgetting rates, and a saturating practice-response gain), and we study when those parameters can actually be recovered from data. For two concepts, we prove a constructive structural identifiability theorem for the associated inverse problem under explicit excitation conditions, together with monotonicity, robustness and L-stability results. We derive an implicit L-stable scheme and a batched solver numerically equivalent to the per-trajectory formulation yet two orders of magnitude faster. The empirical study is two-sided. Under the theorem's excitation conditions, synthetic recovery is exact. On large observational benchmarks it is not. An apparently strong recovery, with forgetting rates correlating with topic error rates at Spearman $ρ= 0.83$, is refuted by four controls: it survives destroying the temporal order of the data, is matched by a classical Bayesian baseline, and is unaffected by removing real timestamps. A single shared forgetting rate attains the same accuracy, and replaying the real schedule with identical true rates reproduces most of the correlation. A positive control shows that the controls detect forgetting where it is observable. We trace the effect to the stationary structure of the model and show that it is consistent with the degeneration anticipated by the theory in the absence of designed excitation.

0.22

PXtal: Learning to Align Powder X-Ray Diffraction and Crystal Structures under Information Asymmetry across Modalities

cs.LG cond-mat.mtrl-sci ·2026-10-07 ·Zhuoran Yang, Christopher M. Collins, Bei Peng, Luke M. Daniels, Matthew J. Rosseinsky +1 more
Abstract · 22 pages, 2 figures, 14 tables

Scientific multimodal learning commonly assumes that paired views are comparably informative. Powder X-ray diffraction (PXRD) makes this mismatch explicit: compressing a three-dimensional crystal structure into a one-dimensional diffraction pattern loses information and makes the pattern harder to connect to the crystal structure that produced it. We introduce PXtal, a framework for learning aligned PXRD and crystal representations under this physically imposed information asymmetry. PXtal uses Unbalanced Optimal Transport (UOT) to adapt the cross-modal coupling and coupling-level generalized Kullback-Leibler (GKL) divergence to supervise the full transport plan. Across six test sets, including four zero-shot transfer sets, PXtal consistently outperforms the baseline models in PXRD-to-crystal candidate retrieval, with the largest gains when PXRD patterns have close but crystallographically distinct nonpaired neighbors, meaning similar input patterns associated with different crystals. The resulting crystal and PXRD encoders transfer more effectively to downstream materials and crystallographic tasks. These results identify information asymmetry as a general design problem in scientific multimodal learning: alignment objectives should reflect what each modality preserves.

0.22

Stale, Misattributed, or Late: Where Personal Memory Fails Before Generation

cs.AI ·2026-10-07 ·Haonan Deng, Park Sinchaisri
benchmark
Abstract

Personal memory for language agents is usually judged by whether the final an- swer is correct. That score hides errors that arise before generation: the memory block may contain an obsolete value, a fact about the wrong person, or no use- ful fact before the serving deadline. We measure these failures directly. Using Personal Fact Memory (PFM) as a reference layer, we find that temporal validity is primarily a property of memory construction in our setting. On a controlled revision benchmark, serving only the active value of each correctly keyed slot eliminates observed stale exposure; without update resolution, 70.3% of prompts expose a superseded value. Once retrievers share the same active store and par- ticipant information, participant-aware BM25 is equivalent to the reference ranker within a prespecified 0.02 margin. The harder problem is assigning revisions to the right slot. Missed merges leave stale values active, whereas false merges silently remove current values; four LLM key assigners achieve higher key re- call than a rule extractor yet produce lower clean-retrieval rates, and open-domain merge recall on LongMemEval never exceeds 0.062. Misattribution survives va- lidity filtering: an entity posterior reduces same-name exposure on controlled data but cannot distinguish identically named speakers in LoCoMo. Two frozen lan- guage models reproduce prompt errors in generated text. Retrieval latency varies across rankers, but prompt prefill dominates turn-level latency on our hardware. These results argue for evaluating agent memory before generation, separating stored-state validity, identity resolution, abstention, and serving latency.

0.22

GeoPrior-Mamba: Structured Process Priors with Mamba for Fine-Resolution XCO2 Reconstruction

cs.LG cs.AI ·2026-10-07 ·Zhao Meng, Yinan Cai, Siru Zhong, Juepeng Zheng, Haohuan Fu
Abstract

Reconstructing fine-resolution column-averaged dry-air CO2 (XCO2) fields from sparse satellite observations requires models to infer spatial structure that is only weakly constrained by direct measurements. Existing learning-based methods typically treat environmental covariates as ordinary numerical inputs and must therefore learn heterogeneous source-sink relationships largely from sparse supervision. We introduce GeoPrior-Mamba, a multi-directional Mamba framework augmented with offline language-model-induced structured process priors. Rather than using a language model to predict XCO2, we use it before training to organize relative process knowledge for biospheric uptake, ecosystem respiration, and anthropogenic emissions into deterministic prior tables. These priors are spatially instantiated using geographic, ecological, emission-related, and seasonal information and are adaptively injected into the reconstruction backbone through a lightweight knowledge adapter. Using OCO-2 observations from 2018-2020, GeoPrior-Mamba achieves an RMSE of 0.81 ppm and an R2 of 0.93 on held-out observations, reducing RMSE by 48.2% relative to CAMS background interpolation and by 3.1% relative to Trans-XCO2 under the same evaluation protocol. Ablation experiments show a measurable contribution from the knowledge-prior branch and substantially faster convergence than the knowledge-free Mamba backbone. Independent TCCON evaluation further supports the consistency of the reconstructed fields with ground-based column CO2 measurements. These results suggest that language models can provide a practical mechanism for constructing structured process priors when globally consistent process-response representations are difficult to obtain directly, while remaining outside the numerical prediction loop.

0.22

Matching of signal, noise and hardware timescales for filtering and forecasting of correlated noise signals

cs.LG cond-mat.dis-nnphysics.app-phphysics.data-an ·2026-10-07 ·Joshua Donald, Alex Gabbitas, Arthur G. T. Coveney, Sergey Savel'ev, Pavel Borisov
Abstract

Physical reservoir computing exploits the nonlinear dynamics of physical systems to process time-dependent data with greater energy efficiency than conventional machine learning approaches. However, physical reservoirs have fixed intrinsic response timescales, whereas real-world signals combine deterministic and stochastic components across multiple timescales. Here we show, using a nanoporous niobium oxide reservoir, synthetic noisy signals and cryptocurrency-price volatility, that the relationship among noise correlation time, reservoir memory and forecast horizon determines whether correlated noise is filtered or predicted. Noise varying faster than the relevant reservoir memory and forecast horizon is averaged by the reservoir, whereas the temporal structure of slower-varying noise is sufficient for algorithmic forecasting. We introduce the reservoir memory horizon and forecasting regime index to distinguish these operating regimes. These contributions demonstrate that timescale matching can guide the encoding of input time series and development of physical reservoir architectures that filter, analyse and predict stochastic signal components across distinct temporal scales.

0.21

World Potential Model: Pretrained World Knowledge as Progress Potentials

cs.AI ·2026-10-07 ·Jun Zhao, Jixin Tang, Yang Shu, Jinyang Wu, Yuyang Lu +4 more
Abstract

Long-horizon language agents often receive supervision only from terminal task outcomes, leaving little signal for distinguishing productive intermediate behavior from stagnation or even regression. Rather than learning a separate value function or process reward model for every task, we ask whether pretrained models can recognize task progress from their existing world knowledge. We formalize this capability with a World Potential Model (WPM), a goal-conditioned evaluator of task-relative realized progress in agent contexts. In ALFWorld and ScienceWorld, off-the-shelf pretrained models substantially outperform chance at recovering realized-progress structure without task-specific evaluator fine-tuning. We further anchor these progress judgments to task-specific milestones to obtain scalar world potentials, whose temporal differences provide process-sensitive step-level credit for policy optimization. Under matched comparisons, WPM-guided optimization improves success over outcome-only GRPO across all evaluated configurations. Together, these results provide initial evidence that pretrained world knowledge can support reusable realized-progress evaluation and provide useful supervision for long-horizon agents.

0.21

Boosting and the Expressive Power of Simple Weak Learners via the $γ$-VC Dimension

cs.LG ·2026-10-07 ·Arthur da Cunha, Kasper Green Larsen, Liang-Yu Zou
Abstract

Boosting converts weak hypotheses with a small edge over random guessing into highly accurate predictors, but the expressive power of the resulting classifier can depend strongly on the structure of the base class. We study this phenomenon through the $γ$-VC dimension introduced by Alon et al. (STOC 2021). Our first result shows that this parameter characterizes the sample complexity for weak-to-strong learning up to a constant factor scaling in $γ$. We then sharpen the general relationship between the classic VC dimension and the $γ$-VC dimension. Finally, we also give improved upper and lower bounds on the $γ$-VC dimension for the fundamental concept classes of decision stumps and axis-parallel rectangles in $\mathbb{R}^d$.

0.21

Unrolled Flow Models for Reasoning

cs.LG cs.AI ·2026-10-07 ·Faissal Izermine, Hanru Bai, Oscar Davis, T. Konstantin Rusch
theory
Abstract · 23 pages, 5 figures, 7 tables

Flow matching enables language generation in few steps, but whether additional integration steps improve reasoning remains unclear. We prove that a flow parameterized by a two-layer Transformer can solve graph reachability, with the required number of integration steps increasing with the target's distance from the root. Yet, standard flow language models can fail to benefit from additional steps on reasoning tasks. We attribute this limitation to objectives that supervise each time point independently, without explicitly training successive steps to build on one another. To address this, we instead train through the model's own latent rollout over a randomly sampled subinterval of [0, 1], decoding only at the endpoint. On ProsQA, this raises accuracy to 97% and enables performance to improve with additional integration steps. For the longer rollouts required by reasoning tasks such as Sudoku and Maze, retracting the latent state onto a sphere stabilizes the dynamics and yields substantial gains over baselines with more than three times as many parameters. Sampling multiple rollouts further improves performance when paired with a parameter-free selection score, although reliable selection remains challenging for longer answers. Together, these results establish a theoretical basis for reasoning with flows and show how rollout training, stable latent dynamics, and rollout selection help realize this capacity in practice.

0.21

Eigenvalues of the Hessian in Deep Learning: The Origin of Symmetry and Its Breaking

cs.LG math.OCstat.ML ·2026-10-07 ·Yossi Arjevani
Abstract

Hessian spectra at trained models in deep learning exhibit a persistent pattern: eigenvalues organize into distinct clusters, including a large bulk near zero and a few isolated outliers. This paper shows that a natural account of these spectral phenomena emerges when the original setting is understood as a departure from a nearby, otherwise hidden, highly symmetric reference. Modifications, including changes to the architecture, data distribution, or parameter metric, expose a nearby reference configuration whose Hessian exhibits rich invariances-ones not accounted for by weight symmetries. There, symmetry enables a precise description of the spectra, forcing high-dimensional kernels and eigenvalues of large multiplicity. Returning to the original configuration breaks the Hessian symmetry and thereby produces the observed hierarchy of clusters and outliers. The framework is developed in some generality, with a detailed analysis of three-layer ReLU networks and applications to convolutional, graph, and transformer models, as well as to the NTK. The same mechanism is further shown to yield analogous spectral structures in layerwise Hessians and the Gauss-Newton matrix.

0.21

Pack and Extract: An Effective Graph-Based Approach for Feature Selection

cs.LG ·2026-10-07 ·Faisal N. Abu-Khzam, Joseph R. Barr, Cynthia Chaaya, Nadim Obeid, Peter Shaw
Abstract

Many molecular biology data sets contain thousands of features, such as genes in single-cell experiments or other omics data. Standard predictive models often perform poorly in such settings, as they tend to retain a large number of features and fail to adequately handle correlations among them. Combinatorial approaches, such as feature selection based on the Independent Dominating Set (IDS) model, have shown promising results, but appear to have been largely overlooked, possibly due to their high computational complexity. In this paper, we propose an improvement to IDS-based feature selection that (i) favors features exhibiting greater diversity and (ii) generalizes the notion of independent dominating set to a node packing model. The resulting formulation is used as an intermediate step within the feature selection process. We evaluate our approach on five genetic data sets and compare the results with those obtained using existing methods. Our experimental results demonstrate notable improvements in predictive accuracy over traditional models, as well as increased stability of the selected feature sets.

0.21

Coding-Agent Benchmarks Should Match Their Users' Task Flows

cs.LG cs.AIcs.CLcs.SE ·2026-10-07 ·Igor Slinko, Yaroslav Golubev, Sergey Titov
benchmark
Abstract

The evaluation of coding agents generally strives to be as realistic as possible. In our study, we collect 4,782 agent sessions of real software engineers in JetBrains IDEs, which we call Production Sessions. Since our subject is interactive agents, we study the sessions with at least three user messages (33% of the sample). These long sessions differ from issue-derived benchmark tasks in two ways: (i) user requests span a far wider mix of task types - questions about the project's code, planning, review, refactoring, execution - and (ii) users switch between types throughout a session. Long-session samples from three public interaction corpora exhibit markedly different Task Flows (the distributions of session lengths, task types, and type-to-type transitions), so no single interaction distribution is universally realistic: benchmarks should name a target use case and calibrate to measurements from it. We present SWE-TaskFlow, an approach for transforming any issue-derived benchmark: it preserves the verified tasks and tests while steering the interaction toward a target Task Flow through prompt splitting and verifiable repository QA, with a TaskFlow Alignment Score (TFAS) for selecting among generated trajectories. In a pilot on 700 SWE-Bench Pro tasks, solving the task sequentially in several steps approximately doubles agent cost without a stable change in resolve rate: the interaction protocol itself is an important dimension of evaluation.

0.21

Amortized Off-Policy Evaluation for LLMs

cs.LG ·2026-10-07 ·Younwoo Choi, Leo Feng, Vincent Liu, Haanvid Lee
Abstract

Accurate evaluation is central to selecting which LLM to deploy, yet testing a candidate on live traffic exposes real users to an unvetted model. Teams therefore evaluate candidates offline, on data produced by already-deployed models. This is off-policy evaluation (OPE), and it faces two distribution shifts: as a model is updated in post-training, its responses diverge from the logged ones (policy shift), and the reward definition under which it is judged changes with business requirements (reward shift). Classical OPE methods are ill-suited to this continual-deployment setting because they are defined per task and require fitting from scratch on every new logged dataset or reward definition. To address this, we propose PFN-OPE, a prior-data fitted network that amortizes OPE across a distribution of contextual-bandit tasks. We pretrain it once on tasks constructed from a pool of LLM responses scored by several reward functions, in which both shifts occur. At test-time it maps a logged dataset and one sampled target response per prompt to a value estimate in a single forward pass, with no per-task fitting. On HelpSteer2 and UltraFeedback with Qwen, Llama, and Gemma policies, PFN-OPE achieves 2.0 to 9.3 times lower error than the best baselines across all tested configurations in the reward-shifted settings.

0.21

Reading the Room: Foundations, Design, and Challenges of Normative Competence in LLMs

cs.AI cs.MA ·2026-10-07 ·Andrea Wynn, Harsh Satija, Nathan Baek, Anqi Liu, Eric Nalisnick +1 more
Abstract

Human communities are governed by normative systems: shared standards that produce \textit{norms} dictating acceptable behavior, enforced through community sanctioning. Aligning increasingly autonomous AI systems with these norms is a central alignment challenge, complicated by the fact that norms are vast in number, change quickly, and are often arbitrary (e.g., dress or language conventions). Thus, alignment requires \textit{normative competence}: the ability to discern from interaction alone what norms a community enforces without relying on static pretrained knowledge. We introduce a multi-agent community debate setting, where access to debate is governed by synthetic norms, to study normative competence in isolation from pretraining exposure. We show that baseline LLM agents fail to learn norms even when doing so would improve their accuracy. We then experiment with various \textit{normative modules} -- architectural components for norm inference -- finding that norm-following is highly sensitive to both the style of norm and the model powering the normative module, suggesting a lack of generalizability. Furthermore, when idiosyncratic, non-normative behaviors accompany the true norm, LLM agents exhibit an unselective attribution failure: they indiscriminately copy idiosyncratic noise alongside enforced rules, a pattern that persists even when imitating unnecessary behaviors is explicitly penalized. To the best of our knowledge, our work is the first to operationalize and evaluate normative competence in LLMs, demonstrating that current AI systems excel at behavioral mimicry but lack the capacity to discern socially enforced order.

0.21

From Retrieval to Customer Context: Evaluating Frontier-Model Systems for Voice-of-Customer Analysis

cs.LG ·2026-10-07 ·Raviraja G, Viraj Bagal, Prabhath Chellingi
Abstract

Organizations increasingly use frontier language models to analyze customer feedback, but answer quality also depends on how that feedback is organized and made available. We define a \emph{customer context graph} as a unified model of customer and business context. Typed relationships connect customer objects (feedback, conversations, users, and accounts), operational objects (tickets, support agents, opportunities, and competitors), and analytical or action objects (taxonomy concepts, evidence, insights, work items, and outcomes). This lets an agent investigate not only what customers say, but why, who is affected, what action followed, who owns it, and whether it was resolved. For this experiment, the graph is populated from public Cursor feedback; the same architecture can support any type of feedback source. We compare Agentic RAG, a Deep Research Agent, and a Customer Context Graph-backed Agent on the same 9,432 public Cursor feedback records using 30 realistic product, incident, comparison, and metadata questions. Without exhaustive ground truth, we jointly score responses on answer quality (coverage and organization), analytical depth (specificity and decomposition), and evidence quality (citation support and traceability), using a comparative rubric calibrated on 28 of the 30 questions. We sample cited records against their claims and weight the three dimensions equally. Under this aligned rubric, the Customer Context Graph-backed Agent scores 0.961 overall, versus 0.710 for the Deep Research Agent and 0.651 for Agentic RAG, and leads the Deep Research Agent on 27 of 30 paired questions (sign-test p < 10^(-5); strictly best on 26 of 30). Its largest advantage is analytical depth (0.967 versus 0.642), reflecting more specific, hierarchically developed findings with quantified themes and traceable evidence...

0.21

RewardWeaver: Long-Horizon Interactive Learning for Language Agents via Self-Evolving Reward Adaptation

cs.AI ·2026-10-07 ·Hengbo Xiao, Boyao Zhang, Purui Liu, Yuxuan Zheng, Haoran Yin +2 more
benchmark
Abstract · 23 pages, 4 figures

Reinforcement learning with verifiable rewards (RLVR) has driven substantial progress in domains where task outcomes can be reliably evaluated, but long-horizon interaction remains challenging due to sparse terminal feedback and difficult credit assignment. Process rewards provide denser supervision, yet the capabilities most relevant for training can change as the policy evolves: a behavior that is easy to evaluate or frequently deficient need not be the bottleneck currently limiting task success. We introduce RewardWeaver, a self-evolving reward adaptation framework for language agents in long-horizon interaction. RewardWeaver maintains a validated capability space in which the semantics of admitted Rubrics remain fixed, and closes the loop between policy optimization, task evaluation, failure attribution, and reward adaptation. After each training stage, it performs outcome-grounded backward attribution on low-outcome trajectories, aggregates recurrent and policy-controlled capability bottlenecks, and dynamically selects the corresponding process rewards for the next stage. Recurrent failures not covered by the existing capability space trigger a separate, controlled expansion procedure. We evaluate REWARDWEAVER on SOTOPIA, Amazon?HistoryPrice, and a newly constructed Sales Benchmark. Across social interaction, bilateral bargaining, and domain-specific sales, REWARDWEAVER establishes new state-of-the-art (SOTA) results. Ablations further demonstrate the importance of dynamic reward allocation, failure-grounded attribution, and stable semantics for admitted capabilities.

0.21

BoT-GRPO: Efficient Process-Reward RL for Reasoning via Bag-of-Token Aggregation

cs.LG ·2026-10-07 ·Yingxiang Yang, Weihang Xiao, Zhunxuan Wang, Joshua Flashner, Niresh Agarwal
Abstract · Published at the COLM 2026 Workshop on Efficient Reasoning

Reinforcement learning is now central to eliciting reasoning in large language models, while in the popular algorithm Group Relative Policy Optimization (GRPO) every token in a rollout receives the same advantage. We ask how to make process supervision efficient: accelerating convergence and improving final quality without the cost of value networks. We propose Bag-of-Tokens Group Relative Policy Optimization (BoT-GRPO), which extends GRPO to token-level reward models through a length-invariant "bag of tokens" aggregation: it collects all token-level rewards across rollouts, weights each by the inverse of its source sequence length, and computes per-token advantages relative to weighted group statistics. BoT-GRPO is critic-free, and is a drop-in replacement wherever GRPO is used when token-level reward is available. On React front-end code generation, BoT-GRPO reaches $80\%$ compile rate up to $1.9\times$ faster than GRPO and converges faster than modern GRPO variants (GSPO, DAPO, PURE) while reaching higher final compile and VLM-judged win rates. On a second task, AIME mathematical reasoning, BoT-GRPO delivers absolute Pass@$k$ gains up to $8.1\%$ over GRPO in half the steps. For both tasks we compare the algorithm's performance on reasoning vs. non-reasoning base-model families (Qwen2.5-3B, SmolLM3-3B, Phi-4-mini-reasoning). Our experiments also yield a practical recipe for the reward model itself: reward stability matters more than richness: clean, bounded, stable fine-grained signals consistently accelerate learning where noisier alternatives stall.

0.21

The Handover Problem: Governing Autonomy Transitions in Human-AI Collaboration

cs.AI cs.HC ·2026-10-07 ·Vicente Pelechano, Antoni Mestre, Manoli Albert, Miriam Gil
Abstract · 10 pages, 5 figures, 8 tables. Submitted to IEEE Transactions on Human-Machine Systems

Human-machine systems rarely operate at a fixed level of AI autonomy. As operators and AI systems collaborate over time, control must shift: the AI can take on more responsibility when collaboration is stable, maintain its current role when evidence is ambiguous, or return control to the human when conditions deteriorate. Existing work on adaptive automation, supervisory control, trust in automation, and deskilling explains parts of this problem, but provides no auditable, multi-signal criterion for governing when autonomy should change across multi-cycle workflows. We formalise this challenge as the Handover Problem: deciding, at each operational cycle, whether to escalate, maintain, or revert AI autonomy while keeping the process reversible, recoverable, and auditable. We introduce the Handover Readiness Score (HRS), a transparent composite measure that integrates four signal dimensions: operator readiness, human-AI trust, learning stability, and operational performance. It is combined with a hysteresis-based transition policy that requires sustained positive evidence before increasing autonomy but reverts promptly when conditions worsen. Across software engineering and manufacturing domains, the HRS and hard safety guards address complementary failure regimes: guards enforce immediate corrective action when a single indicator breaches a critical threshold, while the HRS detects the slow, multi-signal erosion of operator readiness that no individual guard can observe. The framework establishes autonomy handover as a governance problem requiring explicit, composite, and auditable criteria. This provides a conceptual and formal foundation that adaptive automation research has not previously provided.

0.20

D-SLR: The Disjoint Row-Sparse plus Low-Rank Decomposition

cs.LG stat.MEstat.ML ·2026-10-07 ·Vincent Szolnoky
theory
Abstract

Compressing a matrix for reconstruction still defaults to the truncated SVD, approximating the data with a single low-rank structure. It is common to reduce the residual further by adding an overlapping row-sparse component, but methods that solve this joint problem often require iterative solvers and tuning of regularization parameters. We propose the Disjoint Row-Sparse plus Low-Rank (D-SLR) decomposition, a closed-form drop-in for the truncated SVD that improves or exactly matches it. D-SLR restricts rows to either being stored verbatim or approximated by the low-rank fit, never both. Under squared error this restriction costs nothing: the joint optimum is attainable disjointly with fewer parameters at every non-trivial rank and stored row count (shape). With zero stored rows D-SLR reduces to the truncated SVD, so it never does worse at equal cost. The algorithm scores the entire error-versus-parameters tradeoff, and the solution is chosen afterwards by a supplied error target or parameter count, or by a selection rule. The grid and solution together cost three SVDs, with no tuning or regularization. We derive an assumption-free, a-posteriori lower bound on the error at every shape, giving each solution a computable certificate on the potential gain of any other choice of rank and stored rows. Experiments on synthetic and real data (LLM embedding tables, network traffic, hyperspectral images) confirm the gains and quantify the certificate.

0.20

Safe on Average, Unsafe in the Tail: When Is the Episodic-Cost Tail Controllable?

cs.LG cs.AI ·2026-10-07 ·Samuel Tetteh, Cody Fleming
Abstract

Safe reinforcement learning seeks policies that maximize return while satisfying constraints on cumulative cost. Most methods impose these constraints on expected episodic cost. Consequently, standard evaluations report mean episodic cost without characterizing how cost is distributed across episodes. A policy that satisfies the mean-cost criterion may therefore remain unsafe in its worst episodes. Mean-cost reporting neither identifies this tail violation nor shows whether it can be brought within budget while preserving return. In this work, we measure the episodic-cost tail using $\mathrm{CVaR}_{0.1}$, the average cost of the worst $10\%$ of episodes. We classify a policy as tail-safe when $\mathrm{CVaR}_{0.1}$ is within the safety budget. This allows us first to identify policies that are safe on average but unsafe in the tail and then to study whether their tail violations can be controlled while preserving return. To identify tail-unsafe policies, we evaluate five standard algorithms on three Safety-Gymnasium navigation tasks. We then examine four constraint families on dense-hazard navigation and assess tail control across four navigation and four locomotion tasks.

0.20

Learning Traffic Flow Dynamics with Stochastic Physics-Informed Neural Cellular Automata

cs.LG ·2026-10-07 ·Federica Bragone, Matthieu Barreau
Abstract · 45 pages, 16 figures

Traffic flow modeling is essential for understanding and predicting the collective dynamics of vehicles on road networks. Cellular automata provide a simple, interpretable yet powerful framework for representing these dynamics via local interaction rules, while retaining the ability to reproduce complex macroscopic traffic phenomena. However, learning local transition rules from data while preserving physically meaningful constraints remains challenging, particularly for stochastic models. In this work, we propose a physics-informed neural cellular automaton (PI-NCA) for data-driven traffic flow modeling. Building on the standard neural cellular automaton (NCA), we design a neural architecture that is physically consistent with the road topology and guarantees conservation of the total number of vehicles, thereby constraining the learned transition rules to physically admissible dynamics. We further extend this framework to stochastic dynamics by parameterizing probabilistic transition rules while preserving the same physics-informed constraints. We evaluate the proposed models on multiple traffic scenarios generated by the well-established Nagel-Schreckenberg and Kerner-Klenov-Wolf cellular automata. The results demonstrate that the PI-NCA successfully learns the dynamics of both traffic models and consistently outperforms a standard NCA, while the stochastic extension captures probabilistic transition rules without compromising the imposed physical constraints.

0.20

PEACE: Covariant learning of nonadiabatic manifolds with parity-resolved Hamiltonians

cs.LG physics.chem-ph ·2026-10-07 ·Rongzhi Gao, Shuguang Chen, Yang Zhou, GuanHua Chen, Ziyang Hu +1 more
Abstract

Nonadiabatic molecular dynamics provides mechanistic insight into light-driven processes and informs the design of molecules and materials for solar energy conversion, photocatalysis and photo switching. Accurately describing these processes requires a representation that respects electronic symmetry and consistently relates energies to interstate couplings. Here we introduce PEACE, which combines a parity-equivariant latent Hamiltonian with a learned electronic connection. Controlled ablations reveal the complementary roles of symmetry-allowed state mixing and electronic-frame variation in reproducing crossing structures and relaxation dynamics. PEACE closely reproduces excited-state population dynamics from first-principles simulations, while its extension to spin-orbit coupling enables simulations of intersystem crossing. These results demonstrate that a more complete incorporation of the underlying physics into learned electronic representations leads to more accurate predictions of nonadiabatic dynamics.

0.20

RT-Safe: Benchmarking Agent Safety in Real-Time Embodied Environment

cs.AI cs.LG ·2026-10-07 ·Tianruo Rose Xu, Jiawei Ren, Yichi Yang, Zhaoxu Zheng, Lianhui Qin
benchmark
Abstract

Rapid progress in AI agents has brought growing attention to agent safety, with extensive evaluation focused on digital environments. As agents move into the physical world, embodied safety becomes increasingly important: failures can cause human injury and costly hardware damage. Beyond selecting safe actions, embodied agents must also operate under real-time constraints: the physical world does not pause while an agent reasons. As pedestrians move and vehicles approach during inference, an action that appears safe at observation time may become unsafe before execution. Real-time embodied safety therefore depends on both decision quality and decision latency. We introduce RT-SAFE, a simulated urban benchmark for evaluating embodied-agent safety under real-time constraints. RT-SAFE combines navigation tasks with moving actors, environmental hazards, and traffic rules, while allowing the world to evolve throughout inference and action execution. Across eight VLMs, agents achieve high task completion yet almost never complete safely: in the hardest setting, only 0.7% of episodes finish without a safety event. More strikingly, matched static and real-time evaluations yield task completion rates of 91.3% and 94.1%, respectively, while real-time execution increases collisions by $12.3\times$. These results reveal that standard task success can mask substantial safety failures, and that decision latency itself can become a source of physical risk. Finally, we show that RT-SAFE can support offline RL training and substantially reduce collision rates while achieving strong task completion.

0.20

SOTA: Stock Options Trading Agents Guided by Option-Implied Return Distributions

cs.AI cs.LGq-fin.PMq-fin.TR ·2026-10-07 ·Yizhen Xie, Mengyang Liu
published
Abstract · Accepted at the NeurIPS 2026 Agenthon Workshop

As option markets grow and AI advances, agentic systems for option trading are gaining increasing attention. Language-model-based agents can reason over contextual information such as news, but option trading presents a particularly challenging decision problem: a single stock can have thousands of contracts, and the agent must decide both which contracts to trade and how to combine them. Existing approaches often sidestep this complexity by restricting the policy to a fixed strategy structure, such as a straddle, limiting their ability to switch strategies as market conditions change. We present SOTA (Stock Options Trading Agents), an agentic trading framework for structured option-strategy selection. SOTA abstracts the large option universe into strategy-level decisions while deterministic resolvers handle portfolio implementation. We develop SOTA by post-training Qwen3.8-27B with supervised fine-tuning followed by reinforcement learning. SOTA is evaluated on options on nine large-cap U.S. equities and SPY against rule-based and machine-learning strategy selectors in the same trading environment. Over a six-month out-of-sample period, SOTA earns an 18.3% total return with a Sharpe ratio of 1.60 and a maximum drawdown of 8.96%. We also document an asymmetric role of news: news improves frontier-teacher trajectories, but retaining news during reinforcement learning reduces out-of-sample return from 18.3% to -2.7%.

0.20

Conformal Prediction for Spatially Dependent Data via Sequential Whitening

cs.LG stat.MLstat.APstat.ME ·2026-10-07 ·Ayush Baran Sen, Arkajyoti Saha
Abstract

Split conformal prediction uses prediction errors on held-out (calibration) data to determine how wide the prediction intervals should be. It guarantees distribution-free finite-sample coverage when these errors and the error at the target site are exchangeable. This assumption may fail under spatial dependence and nonrandom sampling geometry. Existing spatial methods use fitting residuals to remove the predictable part of spatial variation from calibration and target errors. However, the spatial variation that only the calibration residuals can predict remains in both the target and calibration errors, reducing the efficiency and stability of the interval. We address this by additionally conditioning on the calibration residuals sequentially, which scales to large networks through nearest-neighbour approximations. Under a correct working covariance and an elliptical residual law, the resulting interval has exact finite-sample coverage under any spatial design, and under further conditions it is asymptotically oracle efficient. We also bound coverage loss under covariance misspecification and develop a diagnostic that identifies regions at risk of undercoverage. In simulated data, our method produces narrower and more stable intervals than global and localized state-of-the-art alternatives. In a national PM2.5 application, it produces narrower intervals within the network and identifies regions at risk of coverage failure.

0.20

Benign Overfitting under Heterogeneous Input Fusion

cs.LG stat.ML ·2026-10-07 ·Houzhen Liu, Xiaobo Xia
Abstract · 51 pages, 6 figures

Benign overfitting is extensively studied when learning from a single high-dimensional input, but its behavior under heterogeneous input fusion remains largely unexplored. We study this question for minimum-norm linear interpolation under a heterogeneous Gaussian design, comparing two statistically dependent input blocks with their fusion while holding the underlying population task fixed. For regression, we identify a full-spectrum covariance certificate whose asymptotic status is independent of the cutoff threshold and prove that it is preserved by every positive-semidefinite joint covariance consistent with the two marginals. This protection is sharp, yet it does not extend to all benign regression problems: outside the certified regime, two benign marginals can have a harmful fusion. For one-sparse Gaussian classification, benignity in the regular regime is characterized by the balance between surviving predictive signal and nuisance contamination. Fusion can move these two quantities in opposite directions, and within this model class every marginal-to-joint benign/non-benign pattern is attainable. We further show that the same fused input can have qualitatively different effects on regression and classification. These results establish that benign overfitting under heterogeneous fusion is determined by the joint signal and spectral geometry created by input interaction, rather than by marginal benignity alone.

0.20

Distillation for Incrimination and Distillation for Capabilities

cs.AI ·2026-10-07 ·Sebastian Prasanna, Jacqueline Tay, Alek Westover
Abstract

Powerful misaligned AI models might recognize alignment evaluations and strategically behave well on them, rendering direct audits uninformative. However, distilling such a model into a weaker benign student places the teacher in a Distillation Double Bind: if misalignment transfers, the student may conceal it less effectively, revealing evidence about the teacher; if it does not, the student may learn useful capabilities while remaining benign. We introduce two distinct distillation approaches, one targeting each outcome. Distillation for Incrimination (DFI) aims to transfer misalignment but not the ability to conceal it. Distilling AuditBench's secret-keeping models into their underlying instruction-tuned model produces students that are significantly more likely than their teachers to admit their hidden behavior when asked, suggesting that knowledge of the behavior transferred more readily than the propensity to conceal it. Confession gains largely disappear when the student does not share the teacher's pretrained base, so DFI should target the teacher's own pre-RL checkpoint, which is weaker than the teacher but shares its base model. Distillation for Capabilities (DFC) aims to transfer capabilities but not misalignment. Among several techniques we evaluate, two are effective: inoculation prompting and training for more epochs on fewer unique examples. Both preserve the capability gains of standard distillation while substantially reducing the subliminal transfer of an animal preference, our proxy for misalignment. Together, these findings demonstrate two ways distillation can be used for AI safety: incriminating misaligned models, and extracting their capabilities without their misalignment.

0.20

Temporal Predictive Multiplicity: Equally Accurate Time Series Models Yield Different Forecast Trajectories

cs.LG cs.AI ·2026-10-07 ·Emanuele Albini, Francesca Toni, Saumitra Mishra, Francesco Leofante
Abstract

Models with near-identical predictive performance can yield substantially different predictions, a phenomenon known as predictive multiplicity. Prior work has mostly studied this at the level of individual scalar outputs. In time-series forecasting, however, predictions across horizons jointly define a trajectory, and horizon-wise comparisons can hide important differences in predictive behavior. To address this problem, we introduce temporal predictive multiplicity, a framework that characterizes disagreement over complete forecast trajectories among models with near-identical predictive performance. We show that constraining predictive performance alone can still admit a broad range of different trajectories. We further show that constraining multiplicity at individual horizons partially reduces, but does not eliminate, trajectory-level multiplicity. Experiments with 19 neural forecasting architectures on 11 datasets confirm that near-optimal models can exhibit substantial variability in the forecast trajectories they produce, and trajectory-level disagreement is largely unrelated to horizon-wise disagreement. Our framework, therefore, exposes a gap in existing multiplicity studies: models with indistinguishable predictive performance imply fundamentally different temporal trajectories, with consequential downstream effects.

0.20

SoftSEEPS improves ML-based precipitation forecasting

cs.LG ·2026-10-07 ·Jost Arndt, Utku Isil, Noelia Otero, Rodrigo Almeida, Wojciech Samek +1 more
Abstract

In this paper we have developed a differentiable approximation of the well-known SEEPS score, which we name SoftSEEPS. This allows the training of a Machine Learning model to forecast precipitation directly. We test SoftSEEPS on the IMERG dataset (0.1 degree resolution) by training a decoder for precipitation on the latent space of a pre-trained low-resolution forecasting model. Combining SoftSEEPS and RMSE in a joint objective is possible with marginal trade-offs in either metric.

0.20

EC-EarthFlow: Probabilistic emulation of daily transient global climate model simulations with flow matching

cs.LG physics.ao-ph ·2026-10-07 ·Kirien Whan, Nikolaj T. Mücke, Karin van der Wiel
Abstract

We introduce EC-EarthFlow, a generative flow matching model that emulates simulations from the physical climate model EC-Earth3. The model is trained on transient simulations from EC-Earth3 (1950-2166, SSP2-4.5) to predict the day ahead temperature field from the previous days temperature as well as annual mean temperature. Predictions are made auto-regressively with rollout periods of between a month and an extended season. Using only this variable of interest, we are able to reproduce the daily variability, spatial patterns, annual cycle and long-term trend from EC-Earth3 at a substantially lower computational cost than the physical model. We demonstrate that EC-EarthFlow is stable for long inference periods, and that it can learn the physical relationships as simulated in EC-Earth3.

0.20

RoBART: Bayesian Additive Regression Trees with Tree-Specific Rotations

cs.LG stat.MLmath.ST ·2026-10-07 ·Jeongung Heo, Seonghyun Jeong
theory
Abstract

Bayesian additive regression trees (BART) can require many splits to approximate boundaries misaligned with the predictor axes. RoBART assigns each tree a rotation shared by all internal nodes, retaining axis-aligned splits in rotated coordinates and constant leaves. We jointly propose a Givens rotation sequence and cutpoints on the resulting grid by Metropolis-Hastings and establish reversibility with respect to the conditional posterior with leaf means integrated out. For additive functions with component-specific rotations and anisotropic Hölder smoothness, we prove posterior contraction in empirical $L_2$ distance and for the noise standard deviation. Under the stated prior, design, and grid conditions, with fixed numbers of predictors, trees, and components and no more components than trees, the rate is a sum of componentwise rates determined by smoothness and the number of rotated coordinates used. We also establish a posterior contraction lower bound showing that there exist functions for which RoBART adapts to the intrinsic dimension but axis-aligned BART does not.

0.20

Rethinking the Tradeoff Between Temporal Encoding and Nonlinear Computation in Spiking Language Models

cs.LG cs.AI ·2026-10-07 ·Hanfei Liu, Shuchang Feng, Yanxia Chen, Changzeng Fu, Shiqi Zhao
Abstract

Spiking language models face a tradeoff between representing continuous semantic features over short temporal windows and retaining costly nonlinear attention operations. We introduce Spora, which jointly designs spike encodings and attention operators. Binary temporal weights let $T$ spikes represent compositional values with up to $T$ bits of capacity, compared with $O(\log_2 T)$ bits for spike-count readout. Unipolar Binary Spiking (UBS) uses thresholds and spike-triggered residual decay to produce non-negative integer codes; Bipolar Binary Spiking (BBS) separates sign and magnitude and learns a scale for signed activations. These representations support accumulation-and-shift dot products and integer-exponent mappings in attention. With four time steps, Spora achieves 76.6 average GLUE score and 44.1 CoLA MCC, improving over SpikeLM by 1.2 and 6.2 points, respectively. Extending BBS to six steps raises these scores to 78.2 and 47.4. Conditional-decay analysis, matched-budget activation-quantization comparisons, event-workload statistics, and fixed-point evaluation further characterize the connection between encoding fidelity and computational cost.

0.19

Kinetic Langevin Meets Split Gibbs: Accelerated Posterior Sampling for Imaging Inverse Problems with Diffusion Priors

cs.LG stat.ML ·2026-10-07 ·Dai Hai Nguyen, Duc Dung Nguyen
theory
Abstract

Split Gibbs sampling (SGS) is a popular framework for posterior sampling in Bayesian imaging inverse problems. It decouples a Gaussian data-fidelity term from a complex prior through an auxiliary variable, so the data variable is updated exactly and only the prior-side conditional is hard to sample. Existing samplers treat this conditional in one of two ways. Plug-and-play SGS runs a multi-step diffusion denoiser at every iteration, which is expensive and lacks non-asymptotic guarantees. Langevin-within-SGS takes cheap overdamped Langevin steps but needs many iterations. We propose RED-KLwSGS, which keeps the exact Gaussian update for the data variable and updates the auxiliary variable with underdamped (kinetic) Langevin diffusions driven by a one-shot denoising score, at the same per-iteration cost as Langevin-within-SGS. We prove non-asymptotic Wasserstein-2 convergence in continuous and discrete time for strongly log-concave priors. We also introduce Joint-RED-KLwSGS, which applies kinetic Langevin diffusions to both variables. Experiments with Denoising diffusion probabilistic models as diffusion priors on FFHQ and ImageNet datasets show faster convergence and high-quality image reconstruction.

0.19

Online Resource Allocation with an Endogenous Markov State: Fewer LP Solves Earn More

cs.LG ·2026-10-07 ·Zhaohua Chen
benchmark
Abstract

We study finite-horizon online resource allocation with i.i.d. requests and an endogenous Markov state on a finite state space: each action affects the transition of the state that governs future rewards and resource consumption. In this problem, a transient fluid LP benchmark upper bounds the expected reward of every nonanticipating policy, while a stationary LP supplies randomized state-dependent controls. We assume that the stationary LP has a unique optimum and identify primal nondegeneracy and irreducibility of the optimal induced kernel as important regularity conditions in this framework. With a known request prior, we show that, under nondegeneracy and irreducibility, both frequent and infrequent re-solving attain $O(1)$ regret. However, under a degenerate optimum, irreducibility yields the sharp worst-case $Θ(\sqrt{T})$ rate for infrequent re-solving, while frequent re-solving can incur $Ω(T)$ regret. Thus, more frequent optimization can perform asymptotically worse. With an unknown request prior, we develop a three-phase U-shaped infrequent re-solving policy that coordinates learning and inventory correction with $O(\log\log T)$ LP solves. When the optimal induced kernel is irreducible and the algorithm is given the optimal target state class and a constant-cost entrance policy, it attains $O(1)$ regret under nondegeneracy and $O(\sqrt{T})$ regret under degeneracy. Without the target-class information, linear minimax regret is unavoidable. Numerical experiments further illustrate the instability of round-by-round re-solving relative to epoch-wise infrequent re-solving, show that thresholding greatly mitigates its loss, and find that infrequent schemes remain dominant under both known and estimated priors.

0.19

Node-level Graph Neural Architecture Search Framework

cs.LG cs.AI ·2026-10-07 ·Lintao Yanga, Sirui Lia, Yaqing Wang, Pietro Liò, Xu Shen +2 more
Abstract

In recent years, Graph Neural Networks (GNNs) and architecture search frameworks have gained extensive application in non-Euclidean data processing, attributable to their superior capacity in managing unstructured data. Nevertheless, traditional approaches typically apply uniform convolution operations to all nodes, regardless of their varying structural and feature characteristics, which can undermine model performance and result in over-smoothing issues as the number of layers increases. To overcome this limitation, in this work, we propose a \textbf{N}ode-Level \textbf{G}raph \textbf{N}eural \textbf{A}rchitecture \textbf{S}earch (N-GNAS) algorithm. It can automatically choose an appropriate network architecture for each subset of nodes when updating node features. N-GNAS also introduces a contrastive learning loss to separate sample features from different categories and vice versa. In experiments conducted on eight datasets for node and graph classification, our methodology outperforms current leading GNAS techniques and traditional human-designed GNNs. For example, it achieves an accuracy rate of 78.26\% on the CiteSeer dataset.

0.19

OrBIT: Structure-Guided Embedding Compression

cs.LG ·2026-10-07 ·Yunied Puig, Amit Kumar Jaiswal
Abstract

Embedding tables are among the largest components of modern language models. Most compression methods fix a coding geometry such as coordinate blocks, low-rank subspaces, or unrestricted codebooks, and optimize within it. We instead ask whether the coding geometry can itself be discovered. We introduce \emph{OrBIT}, a structure-guided embedding compression framework that learns reusable local geometry from orbit dynamics and uses it to constrain a small set of shared codewords. The global reconstruction residual then decides where the fixed coding budget is spent, while redundant overlapping charts let local errors compensate one another after gluing. Our theory shows how tight-chart geometry controls distortion, how the global residual directs sequential allocation, and how data-geometry-guided refinement improves the codec. The resulting orbit machinery is compiled away, leaving a compact decoder in which the learned structure governs what is stored, where capacity is allocated, and how local information is assembled globally. Across four LLM embedding tables, OrBIT achieves $37.9\times$ compression on GPT-2 and over $23\times$ on each 7B table relative to 16-bit storage, while delivering competitive rate-distortion performance against established quantization and low-rank baselines.

0.19

Optimizing Large Language Models with Chained LMOs

cs.LG cs.AI ·2026-10-07 ·Sungyoon Kim, Kaan Ozkara, Youngsuk Park
Abstract

Muon has motivated a growing family of optimizers that compose multiple matrix normalizations, but these methods remain fragmented and lack a unified perspective. We introduce chained linear minimization oracles (chained LMOs), which cast these methods as compositions of LMOs. Despite their empirical success, many chains fall outside the standard LMO framework and can diverge on smooth convex objectives. To explain why composition can nevertheless help, we turn to linear associative memory and show that chaining can improve over Muon under anisotropic embeddings. Empirically, we propose TensorChain, a novel optimizer within the framework that stacks compatible weight matrices across different layers and normalizes the 3d tensor across its axes. In Qwen3 0.6B and 1.7B pretraining, TensorChain outperforms all chained baselines in average token efficiency, with average token savings of 9.6% over Muon at matched validation loss.

0.19

Coverage-Aware Reasoning with Medical Tokens for Diagnosis Prediction

cs.LG cs.AI ·2026-10-07 ·Kaisong Zhang, Haotian Fang, Junmeng Zhou, Hang Lv, Yulan Pan +1 more
code/data
Abstract

Large language models (LLMs) offer promising potential for next-visit diagnosis prediction, owing to their ability to integrate longitudinal clinical evidence and reason over it in natural language. However, reinforcement learning for LLM reasoning commonly rewards each trajectory according to the correctness of its final answer. In next-visit diagnosis prediction, multiple diagnoses can be simultaneously valid, but independently rewarding one diagnosis per trajectory does not distinguish repeated hits from coverage of different diagnoses. The policy can therefore concentrate on a few correct diagnoses, leaving others uncovered. Meanwhile, LLM tokenizers can split ICD codes into several generic tokens with limited clinical meaning, requiring multiple decoding steps to predict each diagnosis and hindering reasoning over a large disease vocabulary. To address both challenges, we propose CARing, a framework that represents diagnoses with compositional Semantic IDs (SIDs) and optimizes reasoning trajectories for multi-label coverage. Concretely, we first encode ontology-enriched disease semantics into compact SIDs through residual quantization, and ground the resulting SID tokens in natural language and longitudinal EHR contexts through multi-task alignment and reasoning-enriched training to unlock transferable LLM reasoning. CARing further improves unordered multi-label prediction through a coverage reward for reinforcement learning and multi-positive supervision. At inference time, the model supports both efficient direct constrained decoding and multi-chain reasoning with rank fusion. On MIMIC-III and MIMIC-IV, CARing exceeds all EHR-trained baselines in weighted F1 and attains the highest top-k recall at every reported cutoff, including R@30 of 46.04% and 46.52% in reasoning mode. Our codes and logs are available at https://github.com/zmlxzyh/CARing-Codes-Logs.

0.19

Stationary Bias and Extrapolation in Nonlinear Two-Timescale Stochastic Approximation

cs.LG cs.AI ·2026-10-07 ·A. Ch. Madhusudanarao, Rahul Singh
Abstract

Constant-step stochastic approximation generally has a nonzero stationary mean error that persists under time averaging. This paper studies that error for nonlinear two-timescale recursions driven by an exogenous finite-state Markov chain. Under stated smoothness assumptions and conditions on the stationary distribution, we derive a first-order bias expansion whose error bound remains uniform as the slow step size becomes much smaller than the fast step size. Fast-manifold coordinates keep the associated covariance equation regular in this limit. For fast step $η$ and slow step $\varepsilon$, the expansion reveals a mixed contribution $\varepsilon^2/η$ alongside terms linear in each step size. This dependence matters for bias reduction: along power-law step-size paths, the bias exponents need not be integers, so Richardson--Romberg extrapolation requires weights matched to the path. An exactly solvable nonlinear Markov example verifies the coefficients. We verify localization for temporal-difference learning and compare finite-run extrapolation at equal update budgets. For finite runs, we bound the initialization error of tail averages on both timescales under an additional coupling assumption. In the special case of additive independent noise, signed third-moment cancellation yields a sharper remainder.

0.19

Plan-and-Patch: Diffusion Language Models for Agentic Planning

cs.AI cs.CLcs.LG ·2026-10-07 ·Syamantak Kumar, Jiang Guo, Hassan Hamad, Hideo Kobayashi, Yi Xiang +4 more
Abstract

Planning is increasingly important for long-horizon agents, where successful execution requires coordinating subgoals, tool use, and intermediate outcomes over many steps. Yet assumptions made during planning may be invalidated by the environment, tools may return unexpected results, or actions may fail. Effective agents must therefore not only generate plans, but also revise them. Such revisions often affect only part of a plan, leaving the preceding and subsequent structure intact. Rather than regenerate the entire plan and risk unnecessary changes, repair can regenerate the affected region conditioned on the preserved prefix and suffix. We introduce Plan-and-Patch, a plan-and-act framework in which a diffusion language model (dLLM) generates a structured, program-like plan through parallel unmasking and repairs it by filling in selected regions while keeping the surrounding steps fixed. We compare DreamReasoner-8B and Qwen3-8B as diffusion and autoregressive (AR) planners. On Natural Plan without task-specific training, diffusion (53.7%) achieves nearly twice the plan repair success rate of AR (27.0%). After task-specific training on agentic benchmarks, ALFWorld and TextCraft, the planners achieve similar observed success in plan generation, while diffusion reduces mean plan-generation latency by 39-46% relative to AR. Our results show that Plan-and-Patch provides a framework for faster plan generation and effective plan repair in long-horizon agents.

0.19

Dual-QK: Sharp Queries and Flat Keys for Prunable 2-bit KV Caches

cs.LG ·2026-10-07 ·Sunjoo Whang, Jungjun Oh, Minsung Kim, Dongho Seo, Jisu Shin +3 more
Abstract

Long inputs and extended generation increase the storage and access costs of the key-value (KV) cache. Low-bit quantization reduces storage and memory traffic, while query-channel pruning can further reduce key-cache reads. Rotation-based quantization redistributes the energy of key outliers across channels. To maintain computational invariance, the same orthogonal transform must be applied to queries, preserving query-key dot products. However, this rotation can disperse query energy, weakening the separation between a few large components to retain and many small ones to prune. We introduce Dual-QK, which uses paired non-orthogonal query and key transforms to address this conflict. Using calibrated query and key statistics, Dual-QK combines partial key whitening with a query-aligned basis to balance key scales for INT2 quantization and concentrate query energy for dynamic channel pruning. Channel-0 protection and bucket-relative RoPE support low-bit accuracy over long contexts. Experiments on four models across five generative benchmarks and long-context retrieval tasks show improved accuracy over OSCAR on most tasks at 40% query-channel sparsity. At a 128K context, Dual-QK provides $6.8\times$ KV-cache compression and an estimated $8.3\times$ reduction in KV read volume relative to unpruned BF16. Under the evaluated configurations, our SGLang implementation achieves up to $3.75\times$ the decoding throughput of unpruned BF16.

0.19

Invariant-Measure Reasoners: Stable Representations for Latent Reasoning

cs.LG ·2026-10-07 ·Yuto Inui, Takuya Konishi, Yoshinobu Kawahara
Abstract

Latent reasoning models repeatedly update a latent state using the same recurrent block. As the recurrent depth increases, the sequence of latent states may converge to a compact subset of the state space without necessarily converging to a fixed point. Existing models typically predict by applying a prediction head to a single latent state. However, the latent state can continue to change even after many updates, potentially making predictions unstable across recurrent depths. To address this instability, we introduce invariant-measure reasoners (ImR), a framework that uses an invariant measure as a stable representation. This measure describes the long-run distribution of latent states on the compact subset and is invariant under updates by the recurrent block. ImR predicts from the expectation of the prediction head's output under this measure. We use ImR in two ways: fine-tuning only the prediction head of existing models and training models from scratch. Both approaches reduce prediction instability and improve accuracy in many settings on maze and Sudoku tasks. In some settings, models trained with ImR exhibit non-fixed-point behavior more frequently than existing models yet achieve high accuracy even with such behavior, unlike existing models. These results suggest that ImR can leverage otherwise destabilizing dynamics for latent reasoning.

0.19

Continual Learning without Continual Training

cs.LG ·2026-10-07 ·Nikita Narayanan, Ritham Majumdarr, Sonali Parbhoo
Abstract

Continual learning requires models to adapt to new domains and new classes while retaining prior knowledge. Many existing methods rely on continued optimization, using regularization, replay, or parameter expansion to prevent new updates from overwriting previously learned knowledge. Instead, we propose replacing continual training with continual inference: a PFN-based model that is meta-trained, and then frozen, adapting to new classes only by extending an in-context evidence set. Our model, Latent Concept PFN, performs in-context Bayesian inference over a latent concept space that captures semantic structure shared across domains and classes. As each new domain or class arrives, exemplars are added to the memory; adaptation reflects updated posterior beliefs over latent concepts rather than gradient updates. No parameters are changed, reducing forgetting. The same method handles both domain and class incremental continual learning without task identity. Concept annotations are only used during meta-training, acting as a soft anchor on the latent space rather than a fixed bottleneck. Unlike fixed-vocabulary concept methods, the model also handles noisy, ambiguous, or incomplete annotations by combining concept labels with raw input evidence to discover distinctions beyond the predefined concept set. Experiments on class and domain incremental learning datasets demonstrate competitive continual learning performance while learning interpretable latent concepts.

0.18

AI Safety Considerations for Agents With Limited Time to Act

cs.AI cs.LG ·2026-10-07 ·Leo Zeitler, Jack Richings, Victoria Nockles
theory
Abstract

In the wake of the increasingly public discussion about AI alignment, recent work has tried to propose specific AI architectures that behave safely. However, the proposed arguments that seemingly demonstrate proved alignment mostly neglect the environment the agent needs to act in. We discuss theoretical bounds for agent-agnostic safety guarantees in environments that can only be partially observed and within which an action is required within limited time. We introduce two realistic scenarios, one with an infinite state space and one with signal mixture. In these scenarios, we prove that even a perfect agent cannot guarantee safe behaviour. It will be argued that for any proof of AI safety or alignment, the environment and associated safe actions need to be specifically considered together with the agent.

0.18

Efficient Best-of-N policy evaluation for inference-time alignment

cs.LG cs.AIstat.ML ·2026-10-07 ·Jonas Schweisthal, Yuxin Wang, Athiya Deviyani, Stefan Feuerriegel, Dennis Frauen
Abstract · Preprint

Best-of-N (BoN) is a common inference-time alignment method that selects the highest-scoring response among N samples from a reference model. Evaluating BoN policies from logged data is challenging under sample-only access because standard off-policy estimators require density ratios that depend on unavailable response likelihoods. In this paper, we propose a sample-only framework for evaluating and selecting BoN policies without access to these likelihoods. We show that the order-statistic structure of BoN allows the required density ratios to be expressed through score-rank probabilities that are estimable from samples alone. We then develop a doubly robust estimator of the BoN policy value (BoN-DR) that efficiently reuses a shared auxiliary sample pool across candidate budgets. We establish valid asymptotic inference even under reward estimator misspecification and prove the efficiency of our BoN-DR estimator. Since larger budgets can amplify errors in the score function and lead to reward overoptimization, we derive two selection rules: (i) maximizing the estimated policy value and (ii) maximizing a lower confidence bound on the improvement over the reference policy, which accounts for estimation uncertainty and provides a no-harm guarantee. Across synthetic experiments and GSM8K with multiple reference and reward models, our framework accurately estimates BoN policy values and selects effective sampling budgets.

0.18

Quantum anomaly detection in real scarce data

cs.LG quant-ph ·2026-10-07 ·Emanuele Casciaro, Fabio Mascherpa, Alfonso Amendola, Filippo Caruso
Abstract

Anomaly detection on small and unbalanced datasets remains very challenging in machine learning, although this scenario is common in several domains, including healthcare, cybersecurity, finance, and energy. Data augmentation and generative AI may mitigate training-data scarcity, but they often fall short because anomalies are, by definition, unpredictable, rare, and highly diverse events compared to high-probability normal data. Overfitting to pseudo-anomalies, model collapse, high-dimensional data, uninterpretable black-box models, and validation challenges are typical issues limiting their practical applicability. In this context, quantum machine learning may provide a promising and more sustainable avenue because it can enable more interpretable models with far fewer trainable parameters and smaller datasets, implementable on energy-efficient quantum hardware. Here, we propose a novel two-step hybrid classical--quantum architecture for sequential data and test it on a realistic scenario in the global energy-transition domain, i.e., automated anomaly detection in large-scale photovoltaic plants. The achieved generalization capability and competitive prediction accuracy may pave the way for new hybrid learning models able to exploit the continuously increasing power of cloud-available and more sustainable quantum accelerators integrated with more traditional energy-hungry High Performance Computing resources.

0.18

Differential Refresh of Materialized Derived State: From a Single-Age Equivalence Limit to an Optimal Per-Partition Allocation, with a Real-System Validation on Materialized Views

cs.LG cs.DC ·2026-10-07 ·Amit Rajula
Abstract

Materialized views, cached aggregates, and trained models are all derived state: query results precomputed from base relations that begin to age the instant they are built, because the data they summarize keeps changing. A common way to keep such state fresh is to replace a fixed recompute cadence with an adaptive trigger -- a weighted staleness score that recomputes when accumulated base-data change crosses a threshold. We show this is the wrong lever and identify the right one. First, an equivalence limit: any refresh trigger that is a static, strictly monotone function of a single shared age of a globally-recomputed artifact is operationally equivalent to a calibrated periodic timer, so a global weighted staleness budget cannot schedule a recompute that a uniform timer could not. The escape is to enlarge the state: refresh partitions (the groups of a GROUP BY, per-tenant views, or data segments) differentially, each on its own schedule. We solve the resulting budget-allocation problem: the optimal per-partition recompute rate is proportional to the square root of partition risk sqrt(w_j*lambda_j), and the policy never costs more than the uniform timer, beating it by a closed-form Cauchy-Schwarz "price of uniformity" that grows with heterogeneity. We validate three ways -- analytic, discrete-event simulation, and a real relational engine (SQLite) maintaining materialized GROUP BY views with real SQL and wall-clock timing -- where the policy lowers realized weighted staleness by 14-15% at matched recompute count, winning on every run, while a naive per-partition exposure threshold does not reliably help. The rule for any periodically-recomputed derived state: a global staleness score is a timer in disguise; the leverage is differential refresh allocated by the square root of partition risk.

0.18

Outperformance Inverse Optimization: Learning Objective Functions that Outperform Agent Decisions

cs.AI math.OCmath.STstat.ML ·2026-10-07 ·Akira Kitaoka
theory
Abstract · 81 pages

Inverse optimization estimates the weights of an objective function that explain observed decisions as optimal solutions, and is used in a variety of fields. For mixed-integer linear programs (MILPs), existing methods aim to reproduce the observations as optimal solutions, and thus learn compromise weights when the observations are suboptimal. We propose outperformance inverse optimization, which instead seeks weights that induce, at each state, an optimal solution outperforming the observed action in every component. We give a loss function that can be evaluated with forward-problem oracles alone and is thus applicable to MILPs, together with gradient-based and DC optimization algorithms for minimizing it. For weights inducing a unique outperforming optimal solution at all observations, we prove that the probability of failing to induce such a solution at a new state (the generalization error) is bounded by a quantity inversely proportional to the number of observations, and that this bound is tight in the number of observations up to logarithmic factors. In experiments on synthetic and real data, the proposed methods improve the prediction of solutions outperforming the actions over existing methods.

0.18

MeshSIPP: Efficient Lattice Planning in Dynamic Environment

cs.AI cs.RO ·2026-10-07 ·Marat Agranovskiy, Konstantin Yakovlev
benchmarktheory
Abstract

Autonomous navigation in dynamic environments requires computing spatiotemporal trajectories that satisfy non-holonomic motion constraints. When the trajectories of the moving obstacles are predictable or known, a promising approach is to rely on the combination of state lattices constructed from precomputed feasible motion primitives and Safe Interval Path Planning -- a search-based algorithm with strong theoretical guarantees. While this approach yields feasible paths, the rich primitive sets needed for smooth navigation induce a large branching factor, which becomes costly when coupled with time-dependent obstacle intervals. To this end, we present MeshSIPP, an efficient planner that removes the computational bottleneck by exploiting the fact that many primitives sweep the same regions and can therefore be validated together. MeshSIPP propagates primitives as spatial bundles, screens them with lightweight bounding-interval checks, and defers the expensive exact departure-time search until a primitive reaches its terminal state. A time-aware pruning rule additionally discards redundant space-time branches early in the search. We prove that the resulting search is complete and optimal. Extensive experiments over more than 6,000 benchmark instances and real-time ROS~2 simulations show that MeshSIPP achieves up to a 3$\times$ speedup over state-of-the-art spatiotemporal planners.

0.18

Efficient Reasoning with Flow Language Models

cs.AI ·2026-10-07 ·Hanru Bai, Faissal Izermine, Oscar Davis, T. Konstantin Rusch
Abstract

Flow Language Models (FLMs) have emerged as a continuous-state alternative to discrete diffusion language models, yet the role of their continuous representations in reasoning remains unclear. We investigate this question by comparing the reasoning efficiency of FLMs and discrete diffusion models, measured by solution accuracy under matched denoising steps. Unlike discrete diffusion, which passes categorical states between denoising steps, FLMs evolve a continuous sequence representation throughout denoising and decodes it into discrete tokens only at the end. Our theoretical analysis shows, from a superposition perspective, how information retained in these continuous states can benefit reasoning. Intermediate-state interventions provide further empirical support for this theoretical account, showing that removing information about alternative candidates reduces subsequent solution recovery. Together, these findings show that FLMs allow evidence for multiple candidates to persist and inform subsequent reasoning before a discrete answer is produced. Furthermore, our experiments on maze planning and Sudoku tasks show that FLMs achieve greater reasoning efficiency in the few-step regime: FLMs achieves higher sequence accuracy than discrete diffusion baselines at matched model sizes and small denoising steps. On maze planning tasks, FLMs can also achieve comparable accuracy with smaller models. For example, on Maze15, FLM reaches the 95\% accuracy target at 64 denoising steps with 36.5\% fewer parameters than MDLM. These findings point to continuous state spaces as a promising foundation for reasoning models that require fewer refinement steps.

0.18

Denoising Blocks, Not Tokens: Efficient Compressed Continuous Diffusion with Branching Token Realization

cs.LG cs.AI ·2026-10-07 ·Xinsong Feng, Peng Du, Zhizhuo Yang, Daniel M. Bikel, Jiayun Wang +1 more
Abstract

Diffusion language models (DLMs) generate text through iterative parallel refinement, offering the potential for higher throughput than autoregressive (AR) decoding. However, most DLMs still maintain one generative state per token, so every denoising step processes a state sequence as long as the output sequence, limiting the throughput gains from parallel generation. Continuous DLMs provide an additional degree of freedom: a single continuous state can represent multiple tokens, allowing diffusion to operate on a much shorter latent sequence. We introduce \emph{Branching Latent Diffusion (BLD)}, which exploits this flexibility by compressing a 1024-token sequence into only 64 block latents, a $16\times$ reduction. BLD combines latent compression with \emph{branching token realization}, where each latent is decoded by a local AR branch and all branches run in parallel. Because strong compression makes joint latent generation difficult, BLD generates the latents in groups, conditioning each group on previously generated latents. In end-to-end evaluation on the same GPU, BLD reduces generation FLOPs by more than $80\times$ and increases throughput by more than $6\times$ relative to the similarly sized ELF-L baseline. Compared with the AR baseline, BLD achieves more than $6\times$ higher throughput and more than $4\times$ lower latency. Despite the compression, BLD maintains competitive local fluency and diversity, although long-range coherence remains challenging. Overall, BLD shows that moving diffusion from token-level states to compressed latent sequences can substantially improve the efficiency of long-sequence generation.

0.18

Why Forget-Only Unlearning Needs Memorization

cs.LG cs.ITstat.ML ·2026-10-07 ·Luka Radić, Vikrant Singhal, Amartya Sanyal
Abstract

Machine unlearning asks for a deletion algorithm whose output is close to retraining from scratch without the selected forget examples. In this work, we study forget-only unlearning, where the deletion algorithm receives only the trained model and the examples to forget, with no retained data or extra training information. We ask whether forget-only unlearning is always possible. We first show that this depends on the learning method: different datasets can produce the same trained model but require very different outputs after the same examples are removed. Using this observation, we derive lower bounds on how accurately unlearning can match retraining and instantiate them for several standard learning algorithms. We then ask what must be true when forget-only unlearning succeeds. To this end, we derive lower bounds on what an algorithm must memorize about the training data to handle arbitrary deletion requests. For simple threshold learners, the required information can be as large as the entire dataset, even though ordinary training keeps only one boundary point. Overall, our results show that information discarded during ordinary learning may be needed later for deletion, so models designed for forget-only unlearning may need to retain more information than standard training does.

0.18

Executing Causal Structure Learning with Linear-Attention Transformers

cs.LG ·2026-10-07 ·Amartya Roy, Sayar Karmakar
benchmark
Abstract · 32 pages, 8 Figures

Transformers can execute algorithms on data given in their input. We ask whether they can do the same for causal discovery. We study a standard continuous method that repeatedly updates a candidate causal graph while enforcing acyclicity. We explicitly construct a fixed-weight transformer whose forward pass exactly reproduces one update of this method, so repeated blocks reproduce its optimization trajectory. The transformer carries the current graph and the algorithm's multiplier between updates. We show that retaining the multiplier is essential for exact execution, since different multiplier values can lead to different next updates. We also give conditions under which, within a fixed stage, the number of updates needed to reach a target accuracy can be computed in advance and rounding errors stay bounded as depth grows. Experiments show that the constructed block agrees with a reference update to floating-point precision, while arithmetic replay on synthetic data and seven published benchmark network topologies inherits the reference solver's successes and failures. This separates accurate algorithm execution from accurate causal recovery. In contrast, the ordinary attention models tested under our training budgets do not reliably execute the update or transfer to larger graphs. Whether gradient training can learn an executor in the architecture class of the construction remains open.

0.18

Unbounded Characteristic and Universal Kernels

cs.LG math.STstat.ML ·2026-10-07 ·Jose Cribeiro-Ramallo, Florian Kalinke, Zoltán Szabó
Abstract

Kernel methods are among the most powerful tools in machine learning and statistics, with a large number of successful applications. Their immense success stems from the flexible function class associated to each kernel---its reproducing kernel Hilbert space (RKHS)---which facilitates statistical analysis, as well as from their computational tractability and applicability to many domains. Multiple notions (such as characteristic, $L_p$-universal, and integrally strictly positive definite) capture the expressivity of kernels and their RKHSs and play a key role in understanding the statistical properties of kernel methods; these concepts and their relations are well-understood for bounded kernels. Even though unbounded kernels have received significant attention over the past decade (for instance, in the construction of kernel-based discrepancy and dependence measures such as the maximum mean discrepancy, the Hilbert-Schmidt independence criterion, and the kernel Stein discrepancy), surprisingly little is known about the relations of these notions in the unbounded case. In the present paper we tackle this severe bottleneck, establishing their relations under mild assumptions.

0.18

SearchWorld: Spatial Value-Grounded Imagination for UAV Object Search via World Models

cs.AI cs.LG ·2026-10-07 ·Yatai Ji, Zhengqiu Zhu, Yong Zhao, Yue Hu, Fanglong Yao +3 more
Abstract · 10 pages,2 figures

Autonomous unmanned aerial vehicle (UAV) object search involves a closed loop of perception, decision-making, and action under partial observability. Urban environments pose several challenges: large search areas and narrow egocentric views limit coverage, dense 3D geometry constrains safe motion, and open-world instructions require identifying a specific target among distractors. Many existing methods mitigate partial observability through explicit maps or memory representations, yet remain largely reactive, reasoning over past observations without explicitly predicting future states. World models enable prospective reasoning through imagined rollouts. However, image-generating world models can incur high inference latency, while spatially grounded planning remains challenging for latent world models. We propose SearchWorld, a recurrent state-space world model that connects explicit spatial memory with value-guided imagination. The model maintains BEV exploration and obstacle memory and decodes a task-aware spatial value layer to guide search. A cognition-action network uses this learned spatial value prior to improve the policy through imagined rollouts, without training a separate scalar critic. Training progresses from world-model learning to expert imitation and imagination-based exploration refinement. On UAV-ON, SearchWorld improves the success rate to 23.8% (19.5% for the strongest published agent) and raises oracle success to 35.5%, while remaining robust on unseen scenes (19.9% success rate). By grounding imagination in explicit spatial representations, SearchWorld enables UAV agents to plan prospectively rather than react.

0.18

LLM-Enabled UAV Dispatch: A System-Level Survey and Taxonomy

cs.AI eess.SY ·2026-10-07 ·Xiao Han, Aoyang Quan, Xiangyu Zhao, Xiangjie Kong, Guojiang Shen
survey
Abstract · Under Review

Unmanned aerial vehicle (UAV) dispatch is beginning to move beyond isolated path planning and optimization-driven resource allocation toward system-level coordination supported by semantic reasoning and LLM-based interfaces. This survey provides a unified characterization of LLM-enabled UAV dispatch systems that bridges semantic intent, symbolic decision-making, and physical UAV execution. Rather than treating LLMs as standalone add-ons, we conceptualize them as a cross-layer semantic orchestration layer connecting human instructions, external solvers, and distributed control modules. We organize the literature into four representative dispatch paradigms: pipeline dispatch, global assignment dispatch, decentralized agentic dispatch, and divide-and-conquer dispatch. For each paradigm, we analyze its decision logic, system structure, control flow, representative methods, and potential LLM roles. We further examine how LLMs support semantic parsing, retrieval-grounded planning, solver orchestration, local agent reasoning, multi-agent coordination, safety assessment, and human-facing explanation. We discuss the implications of these paradigms for scalability, robustness, coordination burden, and verification requirements, and identify open challenges including latency-aware reasoning, grounding reliability, physical feasibility guarantees, edge deployment, privacy protection, and distributed consistency. This survey provides a system-level taxonomy and design perspective for integrating LLMs into safety-critical UAV dispatch systems.

0.18

The Attribution Blind Spot: Layerwise Trajectory Diagnostics for Source Reliance in Retrieval-Augmented Language Models

cs.AI ·2026-10-07 ·Zhe Yu, Wenpeng Xing, Yunzhao Wei, Bo Yang, Chen Ye +2 more
Abstract · 27 pages, 4 figures

A retrieval-augmented model can match a document without relying on it. Controlled knowledge conflicts make source choice observable and let us ask a second question that prediction alone cannot answer: which internal-state properties define useful intervention directions? We study paired hidden-state changes with Latent Trajectory Shift (LTS), a signed projection onto a training-fitted first principal component (PC1), and keep verified training exposure separate from behavioral source choice. Across the evaluated conflicts, state-change magnitude is often the stronger predictor, whereas signed PC1 is the stronger selective controller: equal-norm interventions change source preference while better preserving non-target behavior, and the frozen direction transfers across the tested datasets and aligned model pairs. A same-system OLMo study further combines positive choice and control results with inconclusive exposure detection at the achieved power. The central result is a separation: representations that diagnose what a model will choose need not be the representations that best control that choice.

0.17

Q-Learning with Scalar Adjoint Matching

cs.LG cs.AIcs.RO ·2026-10-07 ·Yonghoon Dong, Minsung Yoon, Jaehyuk Kim, Jungwoo Park, Changyeon Kim +1 more
Abstract

Flow policies capture rich and diverse action distributions, and fine-tuning them with off-policy RL to improve beyond the demonstrations has drawn growing interest. However, fine-tuning a flow policy against a learned value function is not trivial, because the policy generates its action over many flow steps. Adjoint matching offers a principled way to update the flow model itself by propagating value information from the final action back to each flow step, but it requires a vector--Jacobian product through the policy at every step, a cost that grows with the number of flow steps and the policy size. We observe that the batch-averaged velocity Jacobian of pretrained flow policies concentrates on its diagonal. Motivated by this finding, we derive a closed-form scalar adjoint that scales the value gradient at the final action by the flow time, eliminating the per-step vector--Jacobian products. We further find that controlling the critic's value at policy-generated actions is particularly important under the scalar adjoint. Based on these findings, we propose Q-learning with Scalar Adjoint Matching (SQAM), which combines the scalar adjoint with a value penalty at those actions. SQAM's gains concentrate on the four hardest OGBench domains, where its success rate exceeds that of the strongest baseline in each domain by 18 to 35 percentage points. To test whether SQAM extends to large pretrained policies, we also fine-tune a vision-language-action policy on a real bimanual robot. SQAM improves over supervised fine-tuning on all three tasks.

0.17

MorphCL: Morphological Contrastive Learning for Inertial-based Human Activity Recognition

cs.LG cs.HC ·2026-10-07 ·Marius Bock, Yuwei Zhang, Juergen Gall, Michael Moeller, Kristof Van Laerhoven +1 more
Abstract

Despite the ubiquity of sensors in wearable and mobile devices and the abundance of human movement data they generate, translating unlabeled recordings into foundational motion models remains an open challenge. Self-supervised learning (SSL) has alleviated the need for costly annotations, yet existing approaches leave the global structure of large-scale motion data largely untapped, relying on randomly sampled batches and local comparisons that become particularly problematic for in-the-wild inertial data dominated by stationary, low-variance behaviors. Here we introduce Morphological Contrastive Learning (MorphCL), a self-supervised pretraining framework that uses structure-aware grouping to inject explicit modeling of global structure into inertial-based SSL approaches. Building on two well-established pillars of motion analysis, the discovery of motion primitives, or motifs, and domain-specific feature descriptors, we show that MorphCL substantially improves linear probing and finetuning results of learned encoders by up to 15 percentage points in F1-score. In a comparison with existing foundation models, we demonstrate that MorphCL-pretrained encoders match or surpass them models in linear probing performance while trained on $4600\times$ less data. Qualitative analysis of the resulting embedding spaces further reveals morphologically meaningful cluster structure, with improved separation of kinematically similar activity classes.

0.17

The Symbol of the Surrogate: Measuring Numerical Provenance in Neural PDE Solvers

cs.LG math.APstat.ML ·2026-10-07 ·Ridham Patel
published
Abstract · Accepted at NeurIPS 2026 AI for Science Workshop

Neural PDE surrogates are trained on numerical solver outputs that contain both physical evolution and solver-specific discretization errors. Because surrogates are also evaluated against held-out trajectories from the same solver, standard benchmarks cannot distinguish fidelity to the exact evolution from imitation of the numerical scheme. We introduce an empirical Fourier-symbol diagnostic that probes a trained surrogate's linearized one-step operator with individual Fourier modes and compares it with both exact-evolution and training-scheme references. To address architectural spectral bias, we train identical networks on schemes with orthogonal dissipative and dispersive signatures and compare their learned operators. In linear advection, the learned surrogates reproduce the training schemes' amplitude and phase errors, with the twin-scheme difference reaching more than 99.8\% of the analytically predicted full-imitation ceiling. The same behavior occurs for a non-local Fourier neural operator and at the operator level for nonlinear Burgers dynamics. These results show that agreement with solver-generated test data does not by itself establish fidelity to the exact evolution. Fourier-symbol measurements provide a direct diagnostic of numerical provenance.

0.17

Decentralized collaborative continual learning: A multi-objective minimization-based technique

cs.MA ·2026-10-07 ·Yara Zgheib, Marc Antonini, Roula Nassif
Abstract

In this work, we formulate decentralized continual learning within a multi-objective optimization framework. For a given inference task t (corresponding to a common minimizer shared by the cost functions of all agents), agents collecting data in a distributed and streamed manner are only allowed to perform local computations and to exchange information with neighboring agents over the underlying communication graph. As tasks evolve sequentially over time, agents must adapt to newly arriving tasks while retaining knowledge acquired from previously learned ones. This requirement leads to the wellknown stability plasticity dilemma, where stability refers to the ability to retain previous knowledge, while plasticity refers to the ability to learn and adapt to new tasks. To address the stability challenge, agents store subsets of samples from past tasks in local memory buffers. Then, through an appropriate multiobjective formulation, the stored information is incorporated into the learning process so that parameter updates account jointly for the current task and previously learned tasks. The proposed decentralized continual learning approach is analyzed in the mean square error sense under general assumptions on the individual cost functions and gradient noise processes. The analysis reveals that cooperation among agents improves the performance of continual learning. In particular, by exchanging information with neighboring agents, decentralized collaborative learning can exploit the diversity of locally observed data and memory buffers to improve the network average mean-square deviation (MSD) across tasks. Finally, simulations illustrate the theoretical findings and the effectiveness of the method in reducing forgetting and improving the average MSD across tasks.

0.17

SECURE-AI: Multi-Signal Uncertainty Quantification for Trustworthy Acoustic Cough Detection

cs.LG cs.AIcs.SDeess.AS ·2026-10-07 ·Sejal Solkhan, Feiya Xiang, Mayur Sanap, Geet Khatri, Edgar Lobaton
Abstract

Acoustic models for cough detection achieve high accuracy on curated benchmarks but fail silently in deployment, where speech, environmental noise, and audio unlike the training data induce confident misclassifications. In health monitoring, a confidently wrong prediction is worse than no answer at all. We present SECURE-AI (Self-aware Estimation of Cough Uncertainty for Reliable Evaluation of AI), a multi-signal uncertainty quantification framework that wraps a frozen MobileNetV3 cough classifier and estimates when its predictions should not be trusted, combining Monte Carlo Dropout, Temperature Scaling, and Mahalanobis distance in the embedding space. On a held-out split of 12,161 Coswara recordings, the probability-space methods are individually insufficient: Monte Carlo Dropout assigns lower uncertainty to incorrect than to correct predictions, and temperature scaling improves calibration but not discrimination. The strongest signal is embedding-space distance: Mahalanobis distance separates true positives from false positives (means 44.7 versus 63.4; Cohen's $d = 0.94$; AUROC 0.86) and, among predicted non-coughs, false negatives from true negatives (AUROC 0.82), so the same signal can flag both false alarms and missed coughs. At a descriptive operating point selected from labeled test data, retaining 90 percent of true-positive detections removes about two-thirds of false positives and raises precision from 59.9 to roughly 80 percent. A rejection threshold tuned on one split does not transfer to another (validation F1 of 82 percent falls to 34 percent on test), so we report threshold-independent AUROC as the main measure. An error taxonomy shows that failures are systematic, with most false positives related to speech. The pipeline runs on Raspberry Pi hardware costing 30 to 90 US dollars. Probabilistic calibration alone is thus inadequate for reliable cough detection.

0.16

Finite-Sample Approximation of Hessian-Guided Perturbed Wasserstein Gradient Flows

cs.LG stat.ML ·2026-10-07 ·Ryotaro Kawata, Atsushi Nitanda, Taiji Suzuki
theory
Abstract

Wasserstein gradient flow extends gradient descent to probability measures. Its Hessian-guided perturbed variant (PWGF) adds Gaussian perturbations to escape saddle points in nonconvex problems. We investigate when its approximation by finitely many interacting particles remains accurate over growing time horizons. Our analysis retains the curvature accumulated along the population-driven reference path: negative curvature can amplify approximation errors, while subsequent positive curvature can damp their influence. This captures favorable scenarios in which temporary instability is compatible with accurate tracking over growing horizons. Under regularity assumptions and a prescribed common perturbation schedule, we prove particle and objective-value tracking bounds on a high-probability event for reference paths satisfying explicit conditions on accumulated curvature. To handle state-dependent Gaussian jumps, we construct a population-first coupling that preserves the reference particles' conditional independence and reduces jump errors to covariance comparison. We verify the conditions in a variance-plus-cosine model, where curvature recovery yields a growing-horizon tracking guarantee. We also establish local attraction, transverse descent, and positive second variation in two regions of a regularized matrix-factorization model, motivating a positive-negative-positive curvature pattern.

0.16

Twist Flow for Inverse Problems

cs.LG physics.geo-ph ·2026-10-07 ·Shiqin Zeng, Zijun Deng, Felix J. Herrmann
Abstract

In Bayesian inverse problems, posterior sampling requires generating samples that are consistent with given observations while capturing the range of plausible solutions. Direct conditional generative models introduce latent noise to model this ambiguity, but paired inverse-problem training can still encourage an almost deterministic map from the observation to the target. As a result, generated samples may be observation-consistent while under-representing posterior variability, especially when the posterior is multimodal, leading to undercoverage, mode distortion, or artificial transitions between distinct feasible solutions. We propose joint twist-flow, an augmented flow-matching formulation that learns a continuous transport from the augmented source state $(z_x, y)$ to the augmented terminal state $(x, z_y)$. Here x is the target variable, $y$ is the observation, $z_x$ is the Gaussian reference coordinate for posterior sampling, and $z_y$ is a Gaussian likelihood-side coordinate associated with the observation branch. Under a Gaussian observation model, $z_y$ is motivated by the normalized observation residual associated with observation compatibility. Its role is not to replace uncertainty in $x$, but to couple generated samples of x to observation consistency, helping reduce likelihood-inconsistent variation while preserving variability in weakly constrained directions. We validate the method on low-dimensional inverse problems with reference posterior samples, where joint twist-flow better preserves multimodal posterior support than a direct conditional-flow baseline. We further evaluate the method on image restoration and seismic subsurface velocity-model inversion, showing increased posterior variability while maintaining observation consistency.

0.16

Evaluating Trajectory Features for Routing Final-Layer Attention

cs.LG cs.AI ·2026-10-07 ·Yupeng Yao
Abstract

Attention routing requires a signal that predicts the value of attention on the current prefix. We evaluate whether hidden-state extrapolation error, curvature and error change improve this prediction beyond uncertainty, one-step displacement, position and state projections. Paired executions of the final attention layer supply signed next-token loss differences in frozen SmolLM3-3B-Base and Qwen3.5-4B-Base checkpoints. Utility-supervised routers are tested on 100 held-out PG-19 books at an identical causal 20 percent invocation quota. None of six prespecified comparisons shows a positive gain after familywise correction. In Qwen3.5, a parameter-matched fixed-projection control lowers NLL by 0.00356 nats/token relative to the trajectory router (95 percent interval 0.00218 to 0.00487). Secondary results depend on the operation removed, feature location and scoring horizon; frozen thresholds also drift substantially at longer horizons. Actual selected-query execution yields small long-sequence latency reductions with increased NLL, while learned routers remain slower during cached continuation. The study identifies limits on the incremental value of these trajectory summaries and separates allocation quality from measured inference benefit.

0.16

A Framework for the Systematic Review of ML Assets in AI Registries

cs.LG ·2026-10-07 ·Alexandra González, Quim Motger, Xavier Franch, Silverio Martínez-Fernández
survey
Abstract

Background: Modern software systems increasingly rely on Machine Learning (ML) assets (i.e., pre-trained models, datasets, benchmarks) for building, evaluating, and integrating ML-based systems. However, current exploration, selection and reuse practices of ML assets are not supported by systematic retrieval methodologies comparable to those used in traditional evidence synthesis. Consequently, in practice, ML asset selection is often presented as a settled design decision, supported by informal justification rather than a traceable, evidence-based, and updatable selection process. Aims: This paper explores how systematic review methods can support ML asset retrieval. In doing so, we aim to make their selection transparent and reproducible, grounded in explicit evidence, and ultimately better suited to its intended use. Method: We analyze established systematic review practices from scientific literature and adapt their phases (i.e., planning, conducting, and documenting) to Artificial Intelligence (AI) registries, treating ML assets as first-class units of analysis. The resulting framework integrates registry-aware search strategies, cross-registry schema alignment, and dependency-driven ML asset exploration. Results: We conceptualize ML asset retrieval as a systematic and reproducible process rather than an ad hoc activity, and propose a framework for structured ML asset discovery. \textbf{Conclusions:} This work illustrates how systematic review principles can be extended beyond scientific literature to support evidence synthesis over evolving AI registries.

0.16

Shared and structured inputs undermine collective random choice by reasoning AI agents

cs.AI physics.soc-ph ·2026-10-07 ·Takahiro Ezaki, Naoto Imura, Katsuhiro Nishinari
Abstract

Random selection is widely used in resource allocation and auditing, making reliable implementation essential for AI-agent systems. Behavioural tests across six reasoning models uncovered threshold and divisibility rules used in identifier-based choices. For threshold-following GPT-6 Sol and Gemini 3.8 Flash, single-agent measurements prospectively predicted correlated participation under shared identifiers and biased participation under distinct identifiers with common timestamp bits. Changing dates, formats and identifier labels revealed when these predictions held. Explicit instructions to randomize independently reduced but did not eliminate shared-input correlation. To test implications for oversight, we asked four models to select customer requests randomly for human review. GPT-6 Sol approached the target rate while selecting predictably from identifiers; the others rarely selected requests. All four closely followed supplied random draws. These findings expose collective and audit vulnerabilities that selection rates alone miss, making input-dependent bias, correlation and predictability central targets for agent evaluation.

0.15

AdaPS-LiNGAM: Adaptive Predecessor Selection for Linear Non-Gaussian Acyclic Models under Small-Sample Settings

cs.LG stat.ML ·2026-10-07 ·Shun Yanashima, Kentaro Kanamori, Hirofumi Suzuki
Abstract

Causal discovery becomes particularly challenging when the available sample size is small relative to the number of variables. This challenge also arises in the linear non-Gaussian acyclic model (LiNGAM), an identifiable framework for causal discovery from observational data. DirectLiNGAM estimates a causal order, which arranges variables so that causes precede their effects, by sequentially identifying an exogenous variable and removing its linear effect from the remaining variables. We establish a structural limitation of this procedure: when the number of variables exceeds the sample size, repeated residualization necessarily becomes degenerate before the full causal order can be determined. Our analysis further reveals that each residual can be reconstructed using only a graph-determined subset of variables already placed earlier in the causal order, termed the active boundary. This result motivates AdaPS-LiNGAM (Adaptive Predecessor Selection LiNGAM), which reconstructs each residual directly from the original observations using an adaptively chosen sparse subset of those earlier variables. The same subset-selection principle is also applied to the final pruning step for edge estimation. Experiments on synthetic data demonstrate that AdaPS-LiNGAM provides accurate causal-structure recovery in sample-limited settings and degrades more gradually as the sample size decreases.

0.15

Noise, Denoise, Correct: MCMC Posterior Sampling with Diffusion Priors in Three Steps

cs.LG ·2026-10-07 ·So Takao, Gregory David Bellchambers, Luke Ye, Sanmitra Ghosh, Michalis Michaelides
Abstract

Pretrained diffusion models are powerful priors for inverse problems, but posterior sampling under nonlinear, non-differentiable forward models remain hard. We introduce diffusion waltz, an MCMC method using SDEdit-style noising-denoising as a proposal, corrected via Metropolis-Hastings for exact posterior sampling without prior evaluation. We further propose injecting observations into the proposal while preserving exactness, using a gradient-free ensemble Kalman update. On a non-differentiable Navier-Stokes initial condition recovery task, diffusion waltz outperforms existing baselines across different noise and nonlinearity regimes.

0.14

Successive Training Stages and Large Language Model Persuasion: Effects of Misalignment, Supervised Fine-Tuning, and Preference Optimization

cs.AI cs.CY ·2026-10-07 ·Antony Dalmiere, Pascal Marchand, Guillaume Auriol, Vincent Nicomette
Abstract

Large language models (LLMs) can be tuned to influence human attitudes, yet the respective contributions of successive post-training stages remain un-clear. This study examines how three successive training stages affect LLM persuasiveness: (1) misalignment through supervised fine-tuning (SFT) on conspiracy data, (2) additional persuasive SFT on argumentative data, and (3) Identity Preference Optimization (IPO), a preference-optimization method. A total of 835 participants recruited on Prolific were randomly assigned to five between-subject conditions (neutral text, conspiracy-trained model, persuasion-trained model, preference-optimized model, and GPT-4) and were exposed to texts on 10 divisive political issues, personalized from their individual profiles in all model conditions. Attitude change was measured as the difference between pre- and post-exposure positions on continuous Likert scales and analyzed with an analysis of covariance (ANCOVA). A significant condition x baseline-attitude interaction, F (4, 825) = 5.33, p < .001, indicated that training effects depended on participants' initial attitudes. Persuasive SFT produced greater attitude change than conspiracy training alone, d = 0.30, whereas IPO provided no additional benefit, d = 0.03, and GPT-4 did not differ from neutral text, d = --0.01. These results show that targeted supervised training on persuasive data increases LLM persuasiveness, whereas preference optimization yields no significant gains beyond it.

0.14

Deep Learning vs. Statistical Models for Multi-Horizon Price Forecasting of Second-Hand Electronics: A Systematic Benchmark

cs.LG q-fin.CP ·2026-10-07 ·Mateusz Buczyński, Michał Woźniak, Konrad Kaczyński, Anna Wróblewska, Sebastian Kuk
benchmark
Abstract · 41 pages, 7 figures, 13 tables

Forecasting resale prices of used electronics is critical for subscription-based platforms where pricing errors translate directly into risk. Unlike structured financial markets, second-hand electronics exhibit high volatility, sparse listing histories, and non-normal price dynamics - yet no systematic time-series benchmark exists for this domain. This paper presents the first multi-horizon benchmark of statistical and deep learning forecasting models for used electronics price prediction. We use a large-scale dataset of daily price listings from Polish online marketplaces (January 2022 to March 2025, 100+ smartphone and laptop models) and evaluate eleven models across six horizons from 1 to 365 days, covering classical methods (ARIMA, ETS, Theta), recurrent and convolutional networks (LSTM, TCN), and modern deep architectures (N-BEATS, N-HiTS, TFT, PatchTST, Informer). Three complementary evaluation protocols assess trajectory fitness, one-shot endpoint accuracy, and cross-horizon transfer. N-BEATS achieves the lowest MAPE beyond 30 days, reaching 8.51% at 365 days versus 14.94% for the best statistical baseline - a 43% reduction. At short horizons (1-7 days), all models converge near 0.72% MAPE and the naive baseline remains competitive. A single N-BEATS model trained at 365 days generalizes to all shorter horizons, eliminating the need for horizon-specific models. N-BEATS and N-HiTS also demonstrate superior hyperparameter stability.

0.14

Factorization-Guided Training and Inference for Diffusion-Drafted Speculative Decoding

cs.LG cs.AI ·2026-10-07 ·Jiakun Fan, Min Xiao, Qiang Wang, Neiwen Ling, Zhichao Liu +3 more
Abstract

Speculative decoding accelerates large language model inference through parallel verification of tokens proposed by a drafter. Diffusion drafters further parallelize proposal generation, but greater proposal capacity or higher mean accepted tokens (MAT) need not yield higher throughput. Because verification accepts only a consecutive prefix, early rejections can render later correct predictions ineffective, limiting MAT gains, while even higher MAT may not improve throughput if per-round latency increases. We introduce DScout, which uses two complementary factorizations to guide diffusion drafter training and inference for higher throughput. For training, DScout factorizes position-wise loss into inclusion, weighting, and prediction loss. This guides the design and combination of mixed-horizon training, intra-block unmasking, and acceptance-weighted loss to align supervision with prefix acceptance and improve MAT. For inference, DScout factorizes throughput as the ratio of MAT to per-round latency, guiding target-conditioning depth, drafter depth, and inference block size selection to identify a high-throughput configuration for each workload. Across six benchmarks, DScout achieves $3.84\times$ and $4.14\times$ mean speedups over autoregressive decoding on Qwen3-4B and Qwen3-8B, respectively, improving mean speedup over the state-of-the-art baseline by 9.1% and 7.1%. Across models and concurrency settings, our inference configuration policy provides up to 9.75% additional mean speedup over DScout's fixed configuration.

0.14

Robust Decentralized Fairness Auditing

cs.LG ·2026-10-07 ·Sayan Biswas, Jade Garcia Bourrée, Anne-Marie Kermarrec, Palak, Martijn de Vos
Abstract

Emerging legislation requires large language models (LLMs) to be audited for compliance with regulatory standards, particularly fairness. Such black-box audits typically assume a single auditor with access to a large, representative set of queries. In practice, it can be difficult for an auditor to obtain such a query set, but multiple auditors can together cover the relevant demographic groups by auditing the LLM collaboratively with their individual query sets. However, relying on multiple auditors raises a fundamental trust problem, as they may act on behalf of the LLM provider to portray a misleading appearance of fairness, i.e., fairwashing. We propose Auditopus, a novel approach for robust decentralized fairness auditing. In Auditopus, auditing proceeds in rounds without a central server. In each round, every auditor issues a fixed number of queries to the LLM, and sends only cumulative statistics vectors of its query results to other auditors instead of sensitive queries in clear. The fairness of the audited LLM is then estimated by aggregating all the vectors. We show theoretically and empirically that even a single adversarial auditor in the network can steer this estimate by fabricating the vectors it sends, making an unfair LLM appear fair. To address this threat, Auditopus has each honest auditor locally down-weight any auditor whose cumulative statistics vectors are statistically inconsistent with previous ones. We implement Auditopus and compare it to robust aggregation baselines on two datasets with two pre-trained LLMs. Against an attacker that optimizes the vectors it sends to make the LLM appear fair, Auditopus reduces audit error by up to 78% on average relative to no defense and at least 62% relative to the robust aggregation baselines. Even when 49% of the auditors are adversarial, Auditopus never lets a very unfair or moderately unfair LLM pass as fair.

0.13

SemanticFold: Latent Sequence Compression SeparatesLanguage Modeling, Decodability, and Reasoning

cs.LG cs.AIcs.CL ·2026-10-07 ·Mingyan Liu, Min Huang
Abstract

We study whether latent sequence compression of prompt prefixes preserves the capabilities that large language models rely on during inference. We introduce SemanticFold, a compression scheme that folds prefix hidden states at learned boundaries, and evaluate it across five model scales: Qwen3-1.7B, Qwen3-8B, SmolLM2-1.7B, Pythia-1.4B, and Pythia-6.9B. We use a fixed-target protocol: a frozen prefix is executed natively or compressed, and both arms teacher-force identical continuation tokens. This design rules out target-selection explanations for likelihood changes. We examine five endpoint families: fixed-target negative log-likelihood, finite-label reasoning accuracy, linear probe accessibility, open-ended generation, and systems-level memory and latency. We find that compression moves these endpoints non-monotonically and that they do not share a single compression threshold. On Qwen3-1.7B at compression ratio R=1.7, compressed-minus-native mean NLL decreases by 0.135 under paired bootstrap with 10000 draws. On SmolLM2 at R=1.2, the mean change is 0.013 higher than native. On both Pythia checkpoints, NLL is effectively unchanged. An NLL decomposition separating sequence shortening from the learned residual transform shows that the favorable Qwen likelihood is attributable primarily to residual adaptation rather than to shortening alone. MLP-only, which applies the transform without shortening, achieves 0.082 lower NLL than Full SemanticFold. Linear probe accuracy and macro AUC change by less than 0.03 in absolute value across conditions, with confidence intervals crossing zero. We conclude that preservation under latent compression has no single scalar certificate: language-model fit, decodability, and reasoning behavior answer different questions and can move in different directions under the same compression operation.

0.13

CircuitGate: Logic-Consistent Circuit-Level Functional Modeling for And-Inverter Graphs

cs.LG ·2026-10-07 ·Qifan Zhang, Ruijie Li, Fangzhou Zhang, Qian Ma, Hui Li +4 more
benchmark
Abstract · 16 pages, 3 figures, 11 tables. Qifan Zhang and Ruijie Li contributed equally. Qian Ma is the corresponding author

And-Inverter Graphs (AIGs) are fundamental representations for logic synthesis and verification in Electronic Design Automation (EDA). As structured representations of complex digital systems, AIGs require models to capture functional dependencies beyond local structure and remain robust to functionality-preserving transformations. In learning-based AIG representation, existing approaches are predominantly based on GNNs and rely on local gate-level message passing, limiting their ability to capture circuit-level functional context and making the learned representations sensitive to topology-specific patterns. Therefore, we propose CircuitGate, a function-aware AIG representation learning framework that advances from gate-level semantics to circuit-level functional modeling. CircuitGate explicitly encodes global primary-input (PI) support and models support-overlap-aware reconvergence between fanins, while incorporating logic-inspired Boolean constraints to encourage functionally consistent representations. We evaluate CircuitGate on the large-scale ForgeEDA benchmark and further validate it on the EPFL and ITC'99 benchmarks. Across equivalent-gate identification and signal-probability prediction tasks, CircuitGate consistently outperforms existing methods, achieving up to 21.7% and 14.2% reductions in MAE, respectively. Under direct ForgeEDA-to-OpenABC transfer without fine-tuning, CircuitGate also achieves the best equivalent-gate identification performance, demonstrating strong cross-dataset generalization. These results demonstrate the effectiveness of modeling circuit-level functional dependencies beyond local topology.

0.13

Cost-Efficient Theorem Proving via Agent Orchestration in Program Verification

cs.AI cs.LOcs.PLcs.SE ·2026-10-07 ·Shuangjie Yao, Nikolaus Holzer, Mark Paul Santolucito, Baishakhi Ray, Suman Jana +1 more
benchmarktheory
Abstract

Program verification establishes software correctness through machine-checkable proofs constructed in theorem provers. It's a guarantee especially valuable for code generated by large language models (LLMs), which is fluent but carries no assurance of correctness. Almost all existing provers, however, pursue pass rates alone at whatever sampling or search budget it takes, and overlook the success-vs-cost frontier; yet real software often carries hundreds of interdependent proof obligations, so what matters at scale is not whether one theorem can be proved, but how many can be proved economically. We introduce CoCo-Prover, which formalizes cost-efficient program proving as metalevel decision-making under cost, grounded on two-level proof graphs: an AND/OR proof hypergraph within each declaration is joined to a lemma-dependency graph across declarations; and at each step, it answers two questions: which open goals to select, and which actions to purchase on these goals. Selection stays symbolic as a topological pass over the proof graphs. Action choice is agent orchestration via metalevel decision-making: an agentic router treats every bounded specialist invocation as a separately priced, best-effort computation, matching heterogeneous specialist agents together with configurations, under evolved routing rules as evidence accumulates. On five program verification benchmarks in Lean 4 including function-level CLEVER, VERINA, and AlgoVeri, and repository-level NTP4VC and Vero, we show that CoCo-Prover achieves a better success-vs-cost frontier than baselines including frontier coding agents and state-of-the-art LLM-based provers: it achieves the best solve rate on every benchmark and up to 100% on two benchmarks. It also reduces cost by up to 30.9% compared to the strongest baseline with the strongest LLM in our evaluation.

0.12

A Proof-of-Concept Study of Weakly Supervised Labeling of Fine-Grained EEG Components for Artifact Attenuation

cs.LG eess.SP ·2026-10-07 ·Lu Wang-Nöth, Hai Huang, Philipp Heiler, Shuqiong Wu, Liyun Zhang +1 more
Abstract

Electroencephalography (EEG) is highly susceptible to electromyographic (EMG) artifacts, whose temporal heterogeneity and spatial-spectral overlap with neural activity can leave mixed sources after blind source separation. Existing artifact-removal methods are further limited by scarce reliable component-level ground truth: expert annotations are costly and subjective, while no established method provides realistic simulation-based ground truth for EMG contamination in multichannel scalp EEG. To address these limitations, we propose a framework combining a frequency-aware high-dimensional representation with Multi-Instance Learning. The representation unfolds separated components into frequency-resolved intra-components, creating a space in which mixed neural and muscular activity becomes more separable, while the weakly supervised learning formulation enables artifact-likelihood scores for individual intra-components to be learned from epoch-level labels without finer-grained ground truth. The resulting intra-component classifier supports fine-grained EMG artifact detection and score-guided attenuation. Experiments on held-out subjects show that the framework learns informative intra-component scores and reduces artifact-related spectral deviations most clearly for jaw tension, with moderate effects for raising eyebrows and limited effects for frowning.

0.12

Force without transmission: a depth-induced rank collapse that no loss on the representation reopens

cs.LG ·2026-10-07 ·Martin Hofmann, Patrick Mäder
Abstract · 16 pages, 8 figures, 2 tables

Training can drive a transformer into a rank collapse: all token representations point in one direction, and learning stops. In a related collapse of attention, a loss term with a bounded corrective force repairs the network during the run. We ask whether such a term repairs rank collapse. We collapse small transformers by weakening their skip connection and treat copies of the collapsed network. No added loss term repaired the collapse, although the stronger kind pushed with about a tenth of the task gradient. The reason was the path, not the strength. The task gradient no longer reached the query and key weights, which decide where attention looks, and the added term's gradient faded before the blocks where the collapse forms. Restoring the skip connection, which changes no weight, reopened this path at once. The rank then recovered, but only far above the scale of collapse. After a burst of high learning rate the path stayed open and the rank recovered untreated. Registered predictions from the path ranked recovery times but did not transfer to this cause. In every case the loss stayed above that of a healthy network after the rank recovered. Whether a collapsed network can be repaired depends on whether the gradient still reaches the weights that must change, not on how strongly a loss term pushes.

0.11

ORDERS: An Empirical Study of Norm-Rank Aggregation for Personalized Federated Learning

cs.LG ·2026-10-07 ·Koffka Khan
Abstract

Personalized federated learning combines shared representations with client-specific predictors, but the contribution of a server weighting rule can be obscured by local training and evaluation choices. We study ORDERS, a configuration that combines a shared backbone, a private residual adapter and classifier, geometric weights assigned by descending update norm, feature alignment, and private-parameter perturbations. The server computes a weighted sum of updates obtained from the same broadcast model; it does not obtain an additional optimization effect from sequential addition. A fully specified evaluation comprises 80 final runs: eight configurations, two datasets, and five training seeds on one fixed partition per dataset. On two-class-per-client CIFAR-10, ORDERS achieves $80.51 \pm 0.79\%$ native mean client accuracy, compared with $79.02 \pm 1.42\%$ for FedPer-R1 and $80.27 \pm 0.73\%$ for the matched uniform-weight control. After common local fine-tuning, the difference from FedPer-R1 narrows to 0.32 percentage points. On Sent140, ORDERS reaches $74.71 \pm 0.49\%$, only 0.69 points above a post hoc client training-majority diagnostic. Ablations provide limited, endpoint-dependent evidence for norm ranking and alignment, and no clear benefit from perturbations. Parameter-payload savings are 5.47% and 0.78%, respectively.

0.11

Ream: Unfolding Mutual Awareness in Human-Agent Workspaces

cs.AI cs.HC ·2026-10-07 ·Peiling Jiang, Sangho Suh, Varsha Kishore, Jonathan Bragg, Haijun Xia +4 more
Abstract

As AI agents work alongside humans in shared workspaces, a mutual awareness challenge arises: agents act at speeds that outpace human monitoring, and users' evolving interests are not always expressed in chat. This challenge is especially pressing in literature review, where both parties retrieve, read, and synthesize a growing body of papers. We present Ream, a literature review workspace that supports mutual awareness through structured artifacts, bidirectional engagement tracking, and localized visualizations. Users can see each party's activity within these documents, and agents can retrieve the same history to guide their work. In studies with eighteen researchers, participants used these traces to inspect evidence, steer agents, communicate through annotations, and reflect on their research focus. Shared histories also helped agents build on earlier work. These findings inform how engagement traces within shared documents can support transparency, personalized assistance, and coordination in human-agent knowledge work.

0.11

Physics-Informed Neural Plasticity: PDE Solvers That Reshape Themselves

cs.LG math.NA ·2026-10-07 ·Chun-Wun Cheng, Bingcheng Hu, Angelica I. Aviles-Rivero
Abstract

Physics-informed neural PDE solvers adapt their parameters to satisfy governing equations, yet their representational structure typically remains fixed throughout training. This rigidity is poorly matched to PDE solutions with strongly heterogeneous complexity across space and space--time, leaving capacity insufficient where the physics is difficult and redundant where it is simple. We introduce physics-informed neural plasticity, a paradigm in which the representation itself reshapes during optimization in response to unresolved physics. We instantiate this principle with Representation Capacity Adaptation for PDEs (ReCAP), a Gaussian-localized solver that dynamically redistributes capacity through local enrichment, residual-directed splitting, gate-based pruning, and function-aware merging. ReCAP uses responsibility-weighted error indicators and the geometry of residual energy to determine where and how to refine. To limit the disturbance introduced by splitting, we introduce quiet-child refinement, which initializes new components by transporting the parent representation while controlling instantaneous functional perturbation. We further establish conditional a posteriori reliability and structural-stability guarantees linking localized physics residuals to solution error and stable refinement. Across five challenging 3D and 4D PDE benchmarks against 11 physics-informed solvers, ReCAP achieves the lowest relative $L^2$ error on every problem, reducing error by $10.7\%$--$27.5\%$ relative to the strongest competing result. These results suggest that physics-informed solvers need not merely learn their parameters---they can learn how their representational capacity should be organized.

0.11

Leakage-Controlled Multimodal Learning for Diagnosis and Progression Prediction in Alzheimer's Disease Research

cs.LG ·2026-10-07 ·Akeem Temitope Otapo, Ghazaleh Khodabandelou, Zuheng Ming, Alice Othmani
Abstract

Alzheimer's disease prediction involves irregular visits, heterogeneous measurements and incomplete modalities. This study presents a multimodal multitask framework combining an adapted SFCN MRI encoder, four causal clinical Transformers, shared fusion and task-specific ODE-GRU dynamics. Fine-tuning and LoRA adapt the final two MRI blocks. Task-DRO balances task losses, while Group-CVaR targets cohort and comorbidity strata. Branch-specific input controls, subject-grouped partitions and empirical causality checks support longitudinal evaluation. Across 2,649 subjects and 17,317 visits from ADNI, OASIS-2 and MIRIAD, internal validation yields diagnosis, stage-1 progression and first-stage-1-visit progression AUROCs of 0.935 +/- 0.002, 0.884 +/- 0.003 and 0.870 +/- 0.005, respectively (mean +/- SD across three seeds). Corresponding hybrid AUROCs are 0.951, 0.909 and 0.896. Next-visit MMSE mean absolute error (MAE) is 1.61 points; worst-stratum diagnosis AUROC is 0.827 +/- 0.008. Sampled ADNI explanations identify task-specific input dependence. OASIS-3 external validation yields network and hybrid diagnosis AUROCs of 0.763 and 0.767, hybrid next-visit progression AUROC of 0.764, diagnosis calibration error decreasing from 0.197 to 0.052, and next-visit MMSE MAE of 0.86. Seed-42 paired ablations of six components yield pooled diagnosis and progression AUROC differences between -0.004 and +0.004; removing clinical encoder inputs lowers diagnosis AUROC by 0.272. The framework integrates longitudinal prediction, missing-modality handling, auxiliary comorbidity modelling and subgroup evaluation within a common pipeline.

0.11

Barron Optimal Transport I: Generative Modeling

cs.LG math.AP ·2026-10-07 ·Evan Dogariu, Joan Bruna
Abstract

Motivated by recent applications in generative modeling and sampling, we introduce a framework for optimal measure transport where cost captures the notion of neural network complexity. In transport-based generative models, samples from a reference distribution (e.g. Gaussian) are mapped to samples of a target distribution along ordinary or stochastic differential equations. These are implemented as deep residual networks when discretized in time, where each hidden layer approximates the associated instantaneous velocity. Thus, given a pair of target and reference measures, a natural question is to search for the most efficient neural representation that implements this transport. Our starting point is the kinetic formulation of OT, due to Benamou and Brenier. We replace the average kinetic $L^2$ energy by the \emph{Barron} energy \cite{bach2017breaking, ma2022barron}, a natural norm which measures the complexity of representing a given vector field with a neural hidden layer, and which captures the adaptive properties of feature learning. This defines a metric on the space of probability measures, complementing existing Wasserstein and Stein geometries. In this work we examine the properties of this metric in the context of generative modeling. As a first application, we quantify the suboptimality of diffusion generative modeling in the Barron geometry by establishing super-polynomial score approximation lower bounds for data generated by neural network pushforwards of the Gaussian. We then investigate the benefit of adaptivity as a way to study alternative generative models. In a companion paper \cite{companionpaper} we leverage the Barron transport geometry for sampling applications, extending the scope of Stein variational gradient methods via feature adaptation.

0.11

Artificial intelligence pathways from weather to climate

cs.AI physics.ao-phcs.LG ·2026-10-07 ·Tom Beucler, J. David Neelin, Hui Su, Shivanshi Asthana, Chris Bretherton +8 more
Abstract · 33 pages, 10 figures. Submitted to "Science Advances"

Deep learning has made rapid advances in weather forecasting: autoregressive models trained on atmospheric reanalyses now rival dynamical models across nowcasting, medium-range, and subseasonal-to-seasonal lead times, producing well-calibrated ensemble forecasts at reduced cost. We review these advances and consider their extension to climate horizons, where the challenge shifts from initial-condition skill to producing reliable statistical responses under altered forcings. AI-powered climate prediction systems must produce credible forced responses to drivers (e.g., greenhouse gases, land-use change) typically outside the observed record. We propose two minimum requirements for AI in climate modeling: (i) external forcing agents must enter explicitly enough to support interventions in which they vary independently; and (ii) robustness must be stress-tested in out-of-distribution regimes, including extremes and counterfactual trajectories. Using leading AI autoregressive emulators and hybrid physics-AI models, we identify development and coupling challenges. Comparing the reported throughput of these models with that of GPU-ported dynamical models highlights how AI can reduce time-to-solution by advancing only the target variables at the required resolution and using longer time steps, rather than integrating a full high-frequency, multivariate state. Diverse AI downscaling strategies can partially substitute for explicit fine-scale resolution, paving the way toward inexpensive local hazard assessment across prediction horizons.

0.10

Why LLM Agents Favor Their Group: Stakes, Observed Norms, and Reputation

cs.AI physics.soc-ph ·2026-10-07 ·Yujiao Chen
Abstract

Language-model agents favor their own group because they have watched their members favor each other. The group label alone does little once the decision has a cost; what drives favoritism is observed behavior, and an individual's own record can override it. We test this in small societies with arbitrary group labels, ten rounds of point sharing, and matched one-shot decisions across fifteen OpenAI models and three Claude models, about 4,400 societies and 3.3 million audited model calls. First, the large effect of a bare group label reported in earlier work appears only when giving others points costs the agent nothing; once the agent can keep points for itself, that effect collapses on every model that shows it. Second, under a stake, interaction history becomes the main source of favoritism: the history effect is statistically positive on 13 of 15 models, reaches about 3.5-8 points out of 10 on 11, grows with the number of rounds played, and extends to labeled strangers the agent has never met. Third, with scripted histories, favoritism falls to near zero under an egalitarian norm and reverses when the agent's own group is seen favoring the other side; stronger models side with an individual's record when it conflicts with the group. Group favoritism is thus conformity to observed group behavior, carried to strangers by the label and overridden by individual reputation. The same account predicts responses to betrayal, scandal, and a free offer to change group: public reprimand repairs betrayal better than apology or restitution, allocation punishment remains confined to the offending member, and a formed group cannot be bought but can, on weaker models, be invited away.

0.10

Gen-PINNs: Generative Adversarial Physics Informed Neural Networks for solving partial differential equations

cs.LG physics.comp-phphysics.flu-dyn ·2026-10-07 ·Muhammad M. Akmal, Kamy Sepehrnoori, Michael J. Pyrcz
Abstract

Physics-Informed Neural Networks (PINNs) are a widely used data-free method for solving Partial Differential Equations (PDEs) using machine learning. With recent advances in Generative Adversarial Networks (GANs), adversarial learning has shown strong capabilities for modeling complex data-driven problems; however, the use of GANs in deterministic physics-informed PDE solutions remains limited. In this work, we first identify limitations of standard PINNs for solving PDEs, including spectral bias, loss imbalance, and optimizer stagnation. We then propose Generative Adversarial Physics-Informed Neural Networks (Gen-PINNs), a unified deterministic residual-adversarial framework designed to improve data-free solutions of PDEs with sharp or shock-front behavior. The generator learns the underlying PDE solution using dynamically weighted physics-informed loss components, while separate discriminators evaluate complementary PDE residual features against ideal zero-residual states. The framework further develops and adapts several methodological components, including a Fourier representation for resolving high-frequency spatial content, an orthonormal spectral diagnostic for quantifying frequency-dependent solution errors, and a modified gradient-based dynamic weighting system for physics, initial-condition, boundary-condition, and adversarial loss objectives. Gen-PINNs is tested against standard PINNs on nonlinear and higher-order PDEs, including the Burgers, Allen-Cahn, and Kuramoto-Sivashinsky equations. The results demonstrate substantial improvements in accuracy and convergence across sharp-front, stiff, and higher-order PDE solutions, highlighting the potential of deterministic residual-adversarial learning as an effective approach for solving challenging nonlinear PDEs.

0.09

Spectrally Targeted Muon

cs.LG ·2026-10-07 ·Vishrut Goyal, Rohan Ramkumar
Abstract

The Muon optimizer orthogonalizes each update matrix, setting all of its singular values to one, and has proven highly effective for training large language models. It remains unclear, however, whether this success comes from amplifying small singular directions that gradient descent neglects or from suppressing large, degenerate directions that disrupt training. We introduce Spectrally Targeted Muon, which orthogonalizes only the singular values above or below a threshold $τ$, so that varying $τ$ interpolates between normalized SGD and Muon. It isolates the relevant singular subspaces with projections computed by Newton-Schulz iteration on a shifted Gram matrix, so no SVD is needed. We evaluate these variants on the CIFAR-10 and NanoGPT speedruns, tracking the effective rank of gradient, update, and weight matrices and a new metric, the alignment of updates with the tangent space of the weight matrix's isospectral manifold. We find three things. First, the small singular values of the momentum are not noise. Orthogonalizing everything except the few largest singular values of each matrix nearly matches Muon while touching only a small fraction of the momentum, whereas orthogonalizing only the top falls well short even though it holds almost all of it. On language models every momentum singular value is far below one, so targeted orthogonalization can only amplify, and Muon wins by making directions that are too small to train on at their raw scale trainable. Second, shrinking the largest singular values is what keeps the parameter spectrum flat; this is cheap, and it is not what drives the loss. Third, AdamW differs from Muon mainly in how slowly it builds structure, which explains its slower start and why warmup helps AdamW but only hurts Muon.

0.09

A Unified Information-Theoretic Approach to Constrained Multi-Fidelity Multi-Objective Bayesian Optimization

cs.LG ·2026-10-07 ·Rikuto Matsumoto, Masanori Ishikura, Masayuki Karasuyama
benchmarktheory
Abstract

Bayesian optimization often involves multiple objectives, constraints, and fidelity levels. We address the challenge of jointly selecting where and at which fidelity to evaluate to identify the highest-fidelity feasible Pareto frontier in this combined setting. From a unified information-theoretic perspective, we measure query utility by the information gain about this frontier, provided by an observation. Since this mutual information is intractable, we derive a variational lower bound using a mixture of under- and over-truncated approximations to the Pareto-consistent region. Multi-fidelity surrogate models propagate the information to arbitrary fidelities, yielding a cost-aware acquisition function without separate heuristics for fidelity selection or constraint handling. Experiments on synthetic, benchmark, and real-world problems demonstrate effectiveness across diverse objective, constraint, and fidelity settings.

0.09

The Cost of Classical Multi-Agent Path Finding

cs.MA ·2026-10-07 ·Alvin Combrink, Sabino Francesco Roselli, Martin Fabian
theory
Abstract · 9 pages, 5 figures

Multi-Agent Path Finding (MAPF) is the problem of planning conflict-free paths for multiple agents in a shared space, each from its start to its goal. Classical MAPF has been the dominant formulation for many years, with its assumptions of discrete time and graph-based conflicts presumably easing the search for solutions. These assumptions limit the physical environments and agents for which a solution is truly collision-free, and also place an upper bound on solution quality that no algorithmic improvements can lift. This work investigates how much solution quality, and in what contexts, the classical MAPF formulation forfeits. Continuous-time MAPF (MAPF$_R$) relaxes these assumptions, making it a natural counter-formulation to compare against across various agent counts and sizes, and graph connectedness, topologies, and resolutions. We find that continuous time and agent shape consideration are worth relatively little on their own; their value comes from enabling an expanded range of move actions, on average improving solution quality by at least $5\%$ on narrow and constrained maps and $17\%$ on maps with open spaces. In some cases, the improvements exceed $20\%$. Doubling the map resolution with classical MAPF recovers less than $3\%$, meaning that little of what is forfeited can be bought back through more compute. This work therefore provides insight on when classical MAPF is a reasonable simplification, and when MAPF$_R$ unlocks significantly higher-quality solutions.

0.08

MORA: Modeling Observed Changes for Drift-Robust Time-Series Anomaly Detection

cs.LG cs.AI ·2026-10-07 ·Xudong Mou, Tiejun Wang, Rui Wang, Hui Wang, Pin Liu +3 more
Abstract

Time-series anomaly detection (TSAD) identifies deviations from patterns learned from historical data. In non-stationary settings, distribution drift and true anomalies can cause similar local changes, making it difficult to tell whether a deviation reflects abnormality or evolving context. Existing methods typically adapt to detected shifts or learn drift-insensitive representations, but do not resolve this ambiguity. We define this problem as \emph{temporal change disambiguation}: determining whether a local deviation is explained by broader temporal evolution. We introduce MORA, a drift-robust TSAD framework that reconstructs the same local target from paired short- and long-term views. The reconstruction gap measures contextual support for a local deviation, and a data-dependent correction mechanism conservatively adjusts the primary local anomaly score. Context can only reduce the score when it improves reconstruction of the same target. MORA needs neither drift annotations nor online adaptation. Experiments on four TSAD benchmarks show strong robustness to non-stationarity while preserving sensitivity to genuine anomalies.

0.08

Efficient Provably Private Classification with a Tabular Foundation Model

cs.LG cs.CR ·2026-10-07 ·Talal Alrawajfeh, Cristiana Diaconu, Ossi Räisä, Sebastian Rodriguez Beltran, Yuan He +3 more
Abstract · 74 pages, 18 figures; includes supplementary information

Tabular data underpin prediction and decision-making in medicine, finance, government and science, but often contain sensitive individual-level information, creating a need for accurate prediction while preserving privacy. Traditional private learning provides formal privacy guarantees, but requires slow dataset-specific optimisation, suffers substantial utility loss under strong privacy, and is often difficult to apply correctly. Tabular foundation models adapt rapidly to new datasets, but existing models lack formal privacy guarantees, and are highly vulnerable to membership-inference attacks, limiting their use on sensitive data. Here we introduce PrivTab, an easy to use tabular foundation model for differentially private classification that embeds a privacy mechanism within its architecture. Pretrained on simulated datasets, PrivTab uses in-context learning to transform sensitive rows into compact, provably private summaries---effectively learning how to learn under privacy. PrivTab outperforms private linear and neural-network baselines under moderate-to-strong privacy, shows negligible membership leakage, maintains well-calibrated predictions under strong privacy, and reduces dataset fitting time by 10,000 times, requiring only a single forward pass. By combining formal privacy, speed, and easy of use, PrivTab brings recent advances in AI to applications where sensitive individual-level data have limited their adoption.

0.07

Neural Sampling with Reweighted Normalizing Flows via the Wasserstein--Fisher--Rao JKO Scheme

cs.LG math.NAmath.PRstat.ML ·2026-10-07 ·Chenguang Duan, Johannes Hertrich, Gabriele Steidl
theory
Abstract

We propose a neural algorithm for sampling from distributions specified by unnormalized Boltzmann densities. Our approach is based on the Jordan--Kinderlehrer--Otto scheme for the Kullback--Leibler divergence in the Wasserstein--Fisher--Rao geometry (WFR JKO scheme). Our contributions are twofold. First, we prove that, for any fixed step size, the exact WFR JKO iterates converge exponentially fast to the target as the number of iterations tends to infinity. Notably, this result requires no structural assumptions on the target, such as log-concavity or a logarithmic Sobolev inequality. Second, we develop a neural implementation of the WFR JKO scheme that parametrizes its transport and reaction components using reweighted normalizing flows. Numerical experiments on challenging multimodal targets demonstrate the promising performance of the proposed method.

0.06

SafeEvo: Deciphering the Safety Alignment Mechanism and Evolution in Language Models

cs.AI ·2026-10-07 ·Miao Yu, Hao Huang, Lu Yuan, Yunpeng Li, Kun Wang +1 more
Abstract

Safety interpretability advances the study of Large Language Model (LLM) alignment from behavioral constraints driven by data or algorithms towards a deeper understanding of internal mechanisms. However, existing works have focused primarily on safety-related representations, attention heads, or neurons after alignment, while largely overlooking the safety mechanisms in pretrained-only models and their evolution across alignment checkpoints. To address this, we propose SafeEvo, an interpretability framework from the circuit (sparse subgraphs of an LLM) perspective. SafeEvo first applies an optimization-based extraction algorithm to identify weak refusal circuits in pretrained base LLMs that can independently express refusal behavior. Causally ablating these circuits completely eliminates the base model's refusal of harmful inputs. SafeEvo then traces the evolution of refusal circuits across successive alignment checkpoints and finds that their structures change progressively, suggesting that the alignment tax may result from refusal-circuit updates affecting utility-related parameters. To validate this, SafeEvo introduces Safety Circuit Alignment (SCA), which confines safety updates to the refusal circuits. Experiments across three LLMs and two alignment algorithms show that, on average, SCA outperforms vanilla alignment in three aspects: \textbf{(1) stronger alignment}, lowering harmfulness score by 63.21\%; \textbf{(2) less over-refusal}, yielding a 58.44\% decrease in refusal rates for benign queries; and \textbf{(3) better utility}, retaining 99.58\% of the original model capabilities.

0.05

NeuralBES: A Differentiable, Control-Aware Emulator for Scalable Building Energy Modeling

cs.LG ·2026-10-07 ·Ting-Yu Dai, Takuya Kurihana, Wing Yee Au, Hon Yung Wong
Abstract · This paper has been withdrawn because it was posted without the required authorization from the authors' employer

Demand-side flexibility i.e. forecasting, shifting, and curtailing residential energy loads, depends on thermal models trusted across millions of heterogeneous buildings. Existing tools force a hard tradeoff: high-fidelity physics simulators such as EnergyPlus are accurate but sequential and require per-building calibration, while purely data-driven sequence models scale but abandon the physical structure that makes their predictions trustworthy. We introduce NeuralBES (Building Energy Simulation), a differentiable emulator that resolves this tradeoff by parameterizing a resistance--capacitance (RC) based thermal model with a shared neural encoder: static building metadata such as floor area, vintage, and HVAC type is mapped to physically bounded capacitances, conductances, and equipment coefficients, which become the coefficients of a scalar linear recurrence solved via a log-space parallel scan, and a predictor--corrector loop closes the thermostat--temperature nonlinearity while preserving full-horizon gradient flow. Trained on the ResStock dataset across three climate zones, NeuralBES handles heterogeneous building archetypes, vintages, and climate zones within a single trained encoder, while black-box baselines produce statistically plausible but physically inconsistent trajectories. On the annual full-year rollout, NeuralBES is the only data-conditioned model that is simultaneously physics-valid and accurate to within 4 MAPE points of the strongest raw-error baseline, while operating at roughly an order of magnitude fewer parameters than the transformer and recurrent baselines; among physics-valid baselines at parameter parity it more than halves the MAPE of the grey-box RC alternative.

Language, Vision & Media 274

Insights from Autoresearch for Solar Panel Segmentation

cs.CV ·2026-10-07 ·Justinas Lekavicius, Kursat Komurcu, Valentas Gruzauskas, Linas Petkevicius
code/data
Abstract · Accepted at AutoML4EO 2026 (non-archival AutoML conference workshop). 4 pages + references. https://automl4eo.org/accepted-papers/

This paper investigates AutoResearch, a protocol in which a coding language model edits a training program under a one-hour GPU budget and retains a change only if validation IoU improves. The protocol is applied to photovoltaic panel segmentation on a frozen real-image split, with DeepLabV3--ResNet-50 held fixed. Three campaigns of 24 experiments, using Gemma~4 12B, Qwen3-8B all improve their one-hour baselines, but retained modifications do not transfer across hardware. The Qwen3-8B configuration, trained on real images only, reaches a test IoU of 0.836 versus 0.833 for the reference GAN-augmented schedule. Research repository https://github.com/VU-AIML/automl4eo-autoresearch-segmentation.

0.68

Decoupling Logic from Persona: Structural Immunity of Edge LLM Agents to Context Pollution

cs.CL cs.AI ·2026-10-07 ·Masaaki Nakatsu, Reno Wang
code/dataunusual
Abstract · 29 pages, 3 figures. v2: adds the public repository and licences (Appendix G), aligns the release statements with it, and corrects reporting…

Edge language-model agents must hold a persona and reason correctly at once, in one context window that fills with history and persona instructions. We study what happens to the logical part of such an agent when that history is long, misleading and persona-heavy (persona-logic interference), and present a Decoupling Architecture (AO-DA) that separates logical inference ("What") from persona expression ("How") into two inference paths on one INT4 base model with hot-swappable LoRA adapters. The logic path receives only the core turn and emits a verifiable structured state (Micro-State); the persona path renders it in character with the full history. In same-base-model ablations on an Apple M2 laptop (Llama-3.1-8B-Instruct and Gemma-3-4B-it, 4-bit; 480 runs: 2 bases x 4 pollution levels x 3 arms x 2 tasks x 2 personas x 5 seeds) we find: (i) the decoupled logic path is structurally invariant to pollution: its prompt stays at 180 (Llama) or 167 (Gemma) tokens while the mixed single-pass prompt grows from 242 to 1,203 on Llama, and its outputs are byte-identical across levels (40/40); (ii) the mixed single pass degrades monotonically (composite logic score 0.669 to 0.150 on Llama, 0.487 to 0.150 on Gemma), mostly by failing to emit the required structured output (80-95% of runs on Llama, 100% on Gemma at the two highest levels); (iii) with the same pollution fed into the logic path, the dedicated-adapter, dedicated-format path is still more robust than the single pass on the 8B model (failure 0-20% vs 25-95% at L2-L4; paired $Δ$ +0.30 to +0.50, Cliff's $δ$ 0.50-0.85, Holm-adjusted $p \le 0.03$) but not on the 4B model, where both collapse. Separation costs one extra decode on a topic's first turn (28.2 s vs 18.2 s on Llama) and buys persona hot-swapping in 1.7 ms (warm median) without re-running the logic path. Code, logic-path adapters, logs and analysis scripts are released.

0.58

Itgan at NADI 2026 shared task: Parameter-Efficient Whisper Adaptation for Robust, Mixed-Dialect and Code-Switched Arabic ASR

cs.CL cs.AIcs.SD ·2026-10-07 ·Ibrahim Almajai
published
Abstract · 12 pages, Arabic NLP 2026 Shared Task

We describe the Itgan systems for the three ASR subtasks of NADI 2026, namely robust country-level ASR (1.1), mixed-dialect ASR (1.2), and Tunisian code-switched ASR (1.3). All three share one recipe, Whisper adapted with LoRA on consumer GPUs, and each was carried by a different addition to it. On 1.1, where the dialect label is given at test time, per-dialect specialists continued from a pooled adapter gave the largest gain, and the submitted system reached 57.1% country-average WER. A post-evaluation linear probe on frozen encoder features routes utterances without the label and recovers 44% of what oracle routing gives. On 1.2 the choice of base model mattered more than adapter capacity, and system combination helped only once we added a decorrelated member, reaching 46.7% WER. On 1.3 our system placed second at 14.49% WER with the lowest CER among the leading submissions, 5.38%. Its last 0.60 WER points came without further training, mostly from an exact weight-space average of independently trained runs, with ROVER voting adding the remainder. Every comparison carries a paired-bootstrap test, and we report eight directions that did not work.

0.58

QuadTok: Quadtree Visual Tokenizer for Autoregressive Image Generation

cs.CV ·2026-10-07 ·Yucheng Mao, Zeyuan Chen, Xiaojun Shan, Xiang Zhang, Divyansh Srivastava +2 more
benchmarkcode/data
Abstract

We introduce QuadTok, a novel framework for visual tokenization and autoregressive image generation. Compared to traditional approaches using 2D grids or 1D token sequences, we propose a hierarchical quadtree structure, bridging the gap between 2D spatial binding and 1D sequence-level flexibility. The QuadTok tokenizer dynamically allocates representational capacity to visually intricate areas while leaving homogeneous regions at a coarse resolution. Compared with a fixed 256-token grid, our ImageNet-trained tokenizer saves approximately 10% of tokens on ImageNet and 9% when transferred zero-shot to the COCO dataset, while maintaining comparable reconstruction fidelity. Furthermore, the natural causality introduced by the tree structure seamlessly enables autoregressive image generation. Conditioned on a quadtree topology supplied before generation, our 947M GPT-style generative model achieves a 2.08 gFID on the ImageNet $256 \times 256$ benchmark. Additionally, leveraging the strong spatial correlation preserved by the quadtree structure, the QuadTok generator enables zero-shot spatially controlled image generation capabilities. Code: https://github.com/myc634/QuadTok.

0.56

When Citations Mislead? A Claim-Level Benchmark for Legal Hallucination Detection

cs.CL cs.AI ·2026-10-07 ·M. Mikail Demir, M. Abdullah Canbaz
publishedbenchmarkcode/data
Abstract · 9 pages, 2 figures, 7 tables. Published in the 21st International Conference on Artificial Intelligence and Law (ICAIL 2026), Singapore. Dat…

Large language models are increasingly used in legal research and drafting, but they can still produce claims that sound convincing without being supported by the cited source. We introduce PARCEL, a benchmark for checking whether a legal claim is supported by the underlying authority. Using recent New York State Court of Appeals decisions, we build a dataset of 3,396 parenthetical-style claims labeled as Supported, Refuted, or Not Found. We cast this task as a three-way natural language inference problem and evaluate several state-of-the-art LLMs in a zero-shot setting. Although the strongest models reach up to 0.97 accuracy, the results also show an important weakness: models still incorrectly mark unsupported claims as supported, even when the full opinion text is provided. Across models, missing support is harder to detect than direct contradiction, and fabricated but plausible citations cause the largest drop in performance. Overall, PARCEL provides a practical benchmark for testing claim-level groundedness in legal RAG systems.

0.55

Alice: A Large-Scale German Benchmark for Rubric-Based Multi-Dimensional Automatic Short Answer Scoring

cs.CL ·2026-10-07 ·Zhifan Sun, Sebastian Gombert, Jannik Lossjew, Tobias Wyrwich, Berrit Katharina Czinczel +4 more
top venuebenchmark
Abstract · EMNLP2026 Main

Automatic Short Answer Scoring (ASAS) is central to NLP for Education. However, openly available benchmarks remain scarce, and existing datasets largely address how well students answer a question directly rather than how well they master underlying concepts (knowledge elements) such as thermal energy or epistemic activities (skills) such as reasoning or claim. To address this gap, we introduce Alice, a large-scale, rubric-based German ASAS dataset that is pedagogically aligned and comprises three subtasks: (i) learning performance (Alice-LP), (ii) knowledge elements (Alice-KE), and (iii) skills (Alice-SK). We further formulate rubric-based ASAS as a rubric-retrieval task and benchmark the dataset with a range of language models, from encoder-only models to lightweight LLMs. We also benchmark the dataset with zero-shot prompting via LLMs and a standard classification baseline. The experiments show that LLMs, in particular, struggle to score knowledge elements and skills in the zero-shot setting. They also indicate that rubric text is often useful, especially for Alice-KE and Alice-SK, while on Alice-LP gains over sample-solution-focused inputs are more modest and vary by model and input format.

0.55

Hard, Yet Reducible: Controlled Forward Transfer for Synthetic Degradation Curation

cs.CV ·2026-10-07 ·Chunming He, Kailai Zhou, Jiaming Zuo, Hanqi Liu, Fengyang Xiao +4 more
Abstract · 17 pages, 4 figures, 9 tables

Selecting synthetic degradations for dense prediction requires an estimate of their training utility, the generalization gain they bring under a finite training budget. Clean and degraded twins share content and labels, suggesting a score based on how much short training reduces the excess error caused by degradation. However, this gap can also shrink when clean performance deteriorates. Measuring the improvement on degraded images alone avoids that confound, but it still credits progress that the same amount of clean training would have produced. We propose the \textbf{controlled Reducible Degradation Gap} (cRDG) for regions defined by degradation type and severity. From a common checkpoint, cRDG runs two budget-matched probes that differ only in one augmentation slot, which holds either a synthetic degradation or a clean augmentation. The score is the gain on held-out degraded images relative to the clean-control probe. Clean harm is a separate feasibility constraint. cRDG reveals a correctable severity band in which training on the degradation yields high controlled gain under the available budget, and the band moves with the predictor, the starting checkpoint, and the training budget. \textbf{Curation of Reducible Bands} (\method) uses cRDG to select synthetic data without changing the predictor. On semantic segmentation and salient object detection, \method{} improves representative predictors under matched synthetic-data budgets and training schedules, extends to existing data-generation pipelines, and preserves clean performance. Code and supporting materials will be publicly released.

0.54

Diagnosing Diversity Collapse and Validating Mask-Conditioned Diffusion for Labeled Microtubule Microscopy

eess.IV ·2026-10-07 ·Mario Koddenbrock, Frederic Rapp, Simone Reber, Erik Rodner
top venuecode/data
Abstract · Accepted at NLDL 2027 (Spotlight) https://huggingface.co/spaces/HTW-KI-Werkstatt/DiffuMT

Mask-conditioned diffusion models have become standard for generating labeled training data in biomedical imaging. But the tools used to evaluate them were built for a different problem. Standard metrics like FID and KID score images through ImageNet-trained features. Those features do not transfer to microscopy, leaving the metrics poorly calibrated for specialized domains. Even with a domain-appropriate feature space, a single blended number can still hide a real failure: a model can produce visually plausible images that still transfer poorly to downstream tasks. We identify a specific failure mode: Late in training, realism keeps improving while texture diversity collapses. The model settles onto a fixed color palette, converging to a high-similarity, low-diversity state that standard metrics do not penalize. Therefore, we propose a diversity-aware diagnostic in a domain-validated feature space. It combines a realism axis based on inter-similarity to real images with a diversity axis based on intra-similarity among generated samples. This diagnostic selects a checkpoint that domain experts cannot reliably distinguish from real recordings in a forced-choice study. It also yields useful downstream segmentation: a segmenter trained only on synthetically labeled data reaches a median skeletonised IoU comparable to one trained on real data, at lower cross-fold variance, and clearly ahead of the best available public alternative in this domain, a parametric renderer. We also reproduce a data-efficient hyperparameter-transfer experiment from prior work, tuning several foundation segmenters on a small labeled subset and evaluating on real images: transfer is stronger with our data. We release DiffuMT on HuggingFace, including the triplet dataset, code to reproduce the downstream-utility validation, and a standalone diagnostic tool for evaluating mask-conditioned diffusion models.

0.54

Mixture of Layers: Dynamic Layer Routing for Visual Reasoning

cs.CV cs.AI ·2026-10-07 ·Jeonghwan Kim, Sofia Stoica, Jiwan Chung, Ansel Blume, Hyeonjeong Ha +3 more
publishedcode/data
Abstract · NeurIPS 2026

Pre-trained vision encoders contain layer-wise visual representations that differ in spatial granularity, semantic abstraction, and sensitivity to local details. However, most Multimodal Large Language Models (MLLMs) rely on only the final or penultimate vision encoder representations or fixed aggregation rules, making visual abstraction largely query-agnostic and limiting access to fine-grained cues such as small objects, spatial details, text, and subtle visual attributes. In this work, we propose Mixture of Layers (MoL), an instruction-conditioned layer routing approach at the visual patch level that dynamically aggregates query-relevant latent representations from intermediate vision encoder layers. Given a text query, MoL predicts routing probabilities over vision encoder layers and performs a top-k sparse aggregation over selected hidden states at either the image level, patch level, or through a hybrid routing mechanism. In doing so, MoL enables query-adaptive access to layer-specific visual features for fine-grained visual reasoning. Our experiments across 7 fine-grained visual reasoning tasks demonstrate substantial performance improvements, especially across fine-grained visual grounding and understanding tasks such as +18.9% improvement on V* in overall accuracy, +4.5% on HRBench4K, and +16.3% on CharXiv compared to the baseline MLLMs, without resorting to multi-resolution inputs, simple interleaving of multiple vision encoders, or increasing the number of patch tokens. We study vision encoders' receptive field scales across different layers and their sampling behaviors to provide an in-depth analysis of why layer-wise sampling is helpful, demonstrating that conditional visual representations are a key step towards better visual perception and reasoning in MLLMs. Our project page is available at https://wjdghks950.github.io/mol.github.io/.

0.54

Bridge Routing Heads: Where Multilingual Multi-hop Reasoning Lives in LLMs

cs.CL ·2026-10-07 ·Seunghan Kim, Minyeong Choe, Hyunil Kim, Haehyun Cho
top venue
Abstract · Accepted at EMNLP 2026

Multilingual LLMs answer the same multi-hop reasoning question across languages, but we lack a mechanistic account of whether they share an internal circuit. We identify Bridge Routing Heads (BRH) in two large multilingual LLMs through a three-stage pipeline. The resulting language-specific head sets exhibit near-complete mutual exclusivity across the five languages, with a mean Jaccard similarity of only 0.017 for Llama 3.1 70B and 0.057 for Qwen 2.5 72B, revealing language-idiosyncratic circuits. Ablating general BRH increases two-hop Negative Log-Likelihood (NLL) by 39-89x the random-head baseline, providing direct causal evidence of their role. Amplifying these heads in a failing target-language pass rescues up to 51.7% of cross-lingual failures, with no training. The two models share this dual-circuit pattern but allocate heads differently: Llama concentrates chaining in a large general pool, while Qwen leans on larger language-specific pools. Together these results show that activation-level intervention alone can recover correct answers from cross-lingual reasoning failures.

0.54

Backdooring Acoustic Foundation Models for Physically Realizable Triggers

cs.SD cs.LG ·2026-10-07 ·Zebin Yun, Eyal Ronen, Mahmood Sharif
published
Abstract · Accepted at RAID 2026

Acoustic foundation models (AFMs) have democratized acoustic applications, enabling powerful models for tasks ranging from speech recognition to speaker verification with minimal resources. However, the security of applications based on AFMs remains largely underexplored. Our work addresses this gap by proposing the Foundation Acoustic model Backdoor (FAB) attack, demonstrating that state-of-the-art AFMs are susceptible to backdooring under practical settings. Despite making minimal assumptions about adversary capabilities (e.g., no access to pre-training data), we show that FAB preserves benign performance while inducing backdoors that survive fine-tuning and cause significant degradation across diverse downstream tasks when activated. Notably, FAB utilizes task-agnostic, physically realizable, inconspicuous, and sync-free triggers (e.g., a background siren). We evaluate FAB using two leading AFMs, nine downstream tasks, and four different triggers. We further demonstrate its effectiveness against established defenses and across both digital and physical domains. While extensive end-to-end fine-tuning can mitigate FAB, such a defense is resource-intensive and task-specific. Our work highlights critical risks to AFMs and calls for advanced defenses.

0.53

WAPR: A Foundation Model for Wide-Angle Refinement in Unseen Object Pose Estimation

cs.CV cs.AIcs.RO ·2026-10-07 ·Yulin Wang, Mengting Hu, Hongli Li, Jianghao Zhou, Chen Luo
top venuecode/data
Abstract · Accepted to ECCV 2026. 19 pages, 4 figures. Yulin Wang and Mengting Hu contributed equally. Corresponding author: Chen Luo

Real-world applications require 6D pose estimation to be accurate, fast, and scalable to unseen objects. This paper introduces WAPR, a zero-shot wide-angle pose refinement model that refines candidate poses with rotational deviations up to 90 degrees. With as few as 12 candidate poses per detected object instance, WAPR supports fast inference within 1 s per frame and reaches a pose-estimation throughput of up to 25 detected object instances per second. To support wide-angle training for rotationally symmetric objects, WAPR uses rotational symmetry priors to canonicalize symmetry-equivalent pose targets before loss computation. We further construct SA6D, a large-scale 6D training dataset with such priors. SA6D obtains KASAL-assisted rotational symmetry priors for 944 GSO scans and expands them through geometry and texture augmentation into about 50K augmented object instances and about 2M rendered RGB-D images. In addition, an angle-balanced loss stabilizes learning across different angular ranges by reducing the influence of uninformative large-error cases. Experiments on seven BOP core datasets show that WAPR achieves state-of-the-art performance in unseen-object 6D pose localization and detection under both fast and unconstrained inference settings. Project page: https://github.com/WangYuLin-SEU/WAPR.

0.52

WorldBench: Evaluating LLMs on Three.js Voxel World Generation

cs.GR cs.AIcs.CLcs.CV+1 ·2026-10-07 ·Krish Bakshi
benchmarkcode/data
Abstract

Large language models can now write complete, interactive 3D worlds as code, but grading those worlds automatically is unreliable. Existing judges take one view of the output: a vision-language model scores a few rendered snapshots, or a language model reads the source. On worlds written by five frontier models we find that the two views disagree on 32% of required items, mostly code that no frame shows, and that fixed views miss small close-up contents. We present WorldBench, a benchmark and judge for open-ended, LLM-generated Three.js worlds. From one prompt describing a floating voxel island with ten biomes, physics, and day/night and seasonal cycles, the judge explores the running world, controlling its clock, orbiting it, and sending a navigator agent to frame each biome, and reads the code for what it sees. Neither channel is trusted on its own: a code quote counts only if it is text the source contains, and visual claims are checked against measured pixels where the property is measurable. A mutation test, in which we remove features by construction, shows that code-only judging gives full credit to four of five removed features, because their code remains in the file. Our judge cuts the points kept on removed features by a third (5.44 to 3.55 of 7.11), and what it still credits is mostly code that exists but never runs. We evaluate five frontier models: Claude Fable 5.1, GPT-6 Astra, Kimi K3, Grok 4.7 and Gemini 3.1 Pro. Code, prompt, tests and judge configuration are available at https://github.com/KrishBakshi/worldbench

0.52

SGF+: Decoupling Gradient Flows for Autoregressive Video Generation

cs.CV ·2026-10-07 ·Zihan Su, Junhao Zhuang, Yaowei Li, Siwen Lu, Haoran Li +8 more
code/datahighly cited
Abstract · Project page: https://zihan-su.github.io/self-gradient-forcing-plus

Autoregressive video generation requires denoising the current frames while writing their key-value representations as context for future predictions. However, these two roles typically share parameters, and we find that their gradients exhibit distinct patterns and systematic negative alignment, hindering the joint optimization of visual quality and temporal consistency. We introduce Self Gradient Forcing Plus (SGF+), which assigns separate parameters to context writing and denoising while preserving their interaction through causal attention. Both roles are jointly optimized using the original generation objective without auxiliary losses, with context writing supervised through its contribution to future predictions. This simple change improves visual quality and long-horizon consistency over the evaluated baselines in both framewise and chunkwise generation, without additional video training data or a longer training horizon. Trained on only 5s rollouts, SGF+ supports continuous generation for up to 24 hours without long-video fine-tuning. These results highlight role-specific parameterization as an effective design principle for high-quality autoregressive video generation and native long-horizon extrapolation.

0.51

Sensitivity as an Arbitrary Output Variable for Differentiable Rendering

cs.GR ·2026-10-07 ·Linas Beresna, Eugene Fiume
top venue
Abstract · 4 pages. SIGGRAPH Asia 2026 Technical Communications

Differentiable renderers expose the derivative of any scalar objective with respect to every scene parameter, yet unlike the primal image, which decades of arbitrary output variables (AOVs) have taught us to decompose, inspect, and composite, these derivatives have no established representation for human inspection. We introduce the sensitivity AOV, a render output carrying the sensitivity of an objective to the scene parameters that influence it. A single reverse-mode pass populates a sensitivity buffer over the scene's parameter hierarchy, from which many views are read rather than re-differentiated, in direct analogy to deferred shading: image-space sensitivity at object and parameter-type granularity, projections onto a freely navigable scene from any inspection viewpoint, and, for spatially varying parameters, per-texel fields carried to the surface through texture coordinates. We separate the fixed camera that defines the objective from the free camera used to inspect the result, and position reverse-mode attribution against its forward-mode dual. Our aim is not a single algorithm but a scaffolding that establishes derivative outputs as first-class render products alongside the primal image.

0.51

OpenEBFM: Energy-based Foundation Models Are Scalable Architecture

cs.CL cs.LG ·2026-10-07 ·Yuan Pu, Yudong Lu, Ting Wang, Zhicong Hong, Qinuo Li +4 more
code/data
Abstract · 12 pages, 7 figures, 3 tables

Energy-based models have shown promising properties for domain-specific tasks but remain challenging in modeling general-purpose scenarios. To bridge this gap, we design energy-steered models (ESMs) for foundational language modeling. ESMs integrate energy steering and token generation within a single architecture to leverage the energy-guided search to refine token generation. Specifically, ESMs first optimize a latent state over the energy landscape, then leverages the refined latent state for token generation. By performing such energy-guided pre-training on large-scale corpora, ESMs learn generally effective energy landscape where favorable tokens correspond to lower energy values. To explore the scaling behavior of ESMs, we pre-train ESMs on three large-scale corpora, then evaluate across 26 downstream tasks spanning a broad range of domains. Our results demonstrate that ESMs have promising scaling behaviors with increasing model size and training data, and outperform GPTs on most downstream tasks. These findings highlight ESMs as a promising architecture for large-scale pre-training and general purpose problem solving. We fully release the model weights and training infrastructure to enable the community to completely reproduce and extend our results. GitHub: https://github.com/datamllab/openebfm HuggingFace: https://huggingface.co/collections/guan-wang/openebfm

0.50

KASALv2: Fully Automatic 3D Rotational Symmetry Classification and Axis Localization

cs.CV cs.AIcs.GR ·2026-10-07 ·Mengxin Zhang, Yulin Wang, Chen Luo, Yongzhe Li, Yijun Zhou
top venuecode/data
Abstract · CVPR 2026. 10 pages, 4 figures. Mengxin Zhang and Yulin Wang contributed equally. Corresponding authors: Chen Luo and Yijun Zhou

Rotational symmetry is an important prior in 6D pose estimation, improving pose accuracy and supporting symmetry-aware evaluation. However, current symmetry annotations for 3D objects remain largely manual or semi-automatic, often requiring predefined types or orders, which limits scalability. This work introduces a fully automatic, reference-free framework for symmetry-type classification, rotational-order identification, and full-axis localization across all eight canonical 3D rotational symmetry types. The method localizes a dominant high-order axis, infers its rotational order through self-consistency analysis, and reconstructs the complete symmetry structure under a hierarchy-guided formulation. A texture-aware extension further models appearance-induced reductions in rotational order while preserving axis orientations. Experiments on idealized and real-world datasets demonstrate strong accuracy and generalization, achieving 94.75% accuracy on 438 symmetric objects in GSO. Training FoundationPose with these priors improves accuracy by up to 0.9% across five BOP datasets, showing that automatically estimated rotational priors improve downstream 6D pose estimation. Code is available at https://github.com/WangYuLin-SEU/KASAL.

0.50

OmniCam: Omni-Camera Trajectory Generation via Geometry-Grounded Pose Token Learning

cs.CV cs.AI ·2026-10-07 ·Zhenyang Liu, Chenjie Cao, Yisu Zhang, Xuhui Zuo, Xiangyang Xue +3 more
Abstract

Camera trajectories control viewpoint changes in video generation, scene reconstruction, and robotic perception. Generating them from language requires both scene geometry and target-aware framing. We introduce OmniCam, an autoregressive model that generates camera pose sequences from a single panorama and textual trajectory descriptions. Its geometry-grounded pose token learning combines three components: a panoramic point-cloud encoder for omnidirectional geometric context; hybrid absolute-rotation and relative-translation tokenization with temporally consistent quaternion signs; and separate geometric and semantic conditioning streams with an explicit 3D target anchor. We also construct OmniCaT, containing 267,700 trajectories across four camera behaviors. On the reported OmniCaT evaluation, OmniCam reduces trajectory errors by 28--47% and collision rate by 65.8% relative to GenDoP retrained on OmniCaT. Against the best baseline for each metric, the ATE and collision reductions are 43.0% and 62.3%, respectively. Component ablations support the use of geometric and target-aware conditioning, while downstream experiments examine camera-controlled video generation and robotic active perception.

0.49

AI4Fire: Evaluating Large Language Models on Wildfire Tasks

cs.CL cs.LG ·2026-10-07 ·Yue Zhao, Zuobin Xiong, Xiyang Hu, Zhangyu Wang, Xiongye Xiao +1 more
code/data
Abstract · 51 pages. Code: https://github.com/yzhao062/AI4Fire

Large language models (LLMs) are entering wildfire management, where overstated evaluations can cost property and lives. How do they perform on wildfire tasks, with and without grounding? Bare means a model receives the task input alone. Grounded means it also receives one task-specific addition: for smoke detection, a smoke-free reference frame from the same camera. AI4Fire runs six core models bare and grounded on five wildfire tasks, zero-shot; a sweep adds 29 more. Our literature search on fire tasks found 138 works; none combines this roster, task coverage, and paired bare and grounded runs. We report three findings. (1) Grounding helped most where the addition carried the answer: a read-only SQL tool lifted every core model's database accuracy from at most 16 to at least 88 percent. (2) Simple rules were hard to beat: no core model outperformed repeating today's staffing count, and two open-weight models mostly copied the median of similar earlier fire-days, a worse forecast. (3) Public releases carry hazards: a fire-danger column separates the holdout perfectly, and 67 aerial fire frames carry smoldering or fire-free labels read from a clipped thermal maximum. We release prompts, responses, scores, code, and the survey record.

0.49

MeshCarve: Artisan Mesh Generation with Flow Matching in Compact Latent Spaces

cs.CV ·2026-10-07 ·Xiyu Wang, Ruocheng Wu, Yufei Wang, Zhihao Li, Lanqing Guo +1 more
hot direction
Abstract · 18 pages, 6 figures, 9 tables

Prior artisan mesh generation works largely predict face tokens autoregressively, which makes inference slow. Recent methods instead flow match continuous latents built by Variational AutoEncoders (VAEs), but reconstruction quality drops significantly when geometry and topology are jointly encoded, and further when the latent space is compressed. We present MeshCarve, a flow matching method that generates entirely in compact latent spaces, generating vertex positions and edge connections separately and sidestepping the difficulty of a joint compact latent. To shorten the token sequence, we propose a hierarchical sparse transformer backbone, instantiated as VertexVAE and EdgeVAE. Instead of encoding fields over the surface voxels, both VAEs anchor on discrete vertices in their latent spaces, which drastically reduces the token sequence length, and our spatial-aware compression shortens it further without costing reconstruction. VertexVAE directly encodes vertex occupancy. For connectivity, we propose vertex-link encoding, which turns arbitrary connectivity between vertices into fixed-length continuous per-vertex embeddings and recovers complex artistic topology faithfully. MeshCarve combines these VAEs with an anchor generator and flow matches on the shortened token sequences. It shows advantages over state-of-the-art autoregressive and flow matching methods on Objaverse and generalizes to Toys4K. To the best of our knowledge, it is among the first artisan mesh generation methods whose every generative stage runs in a spatially compressed latent, with a token sequence only a fraction of the most compressed previous autoregressive and flow matching works.

0.49

Video Prediction Policy 2: Predict Better, Act Better

cs.CV cs.RO ·2026-10-07 ·Yanjiang Guo, Haodong Yan, Zhide Zhong, Zhongru Zhang, Qingyuan Yang +13 more
benchmarkcode/datahighly cited
Abstract · Project Page: https://robert-gyj.github.io/video-prediction-policy-2

World action models (WAMs) have emerged as an important class of generalist robot policies, aiming to transfer video prediction priors to action learning. However, we find that existing WAMs frequently produce incorrect motion predictions in open-ended environment, leading to erroneous actions. We attribute this limitation to two factors: (1) base video models are not optimized for manipulation, and (2) naively incorporating action components into video models can substantially degrade their generalization capabilities. We introduce Video Prediction Policy 2 (VPP2), a WAM that enables strong zero-shot generalization in both video prediction and action generation. First, we curate a large-scale, diverse dataset of manipulation videos to continue pretraining the base video foundation model. We annotate video clips with detailed captions and perform \textit{event-level} video pretraining to promote generalization across open-ended manipulation tasks. Second, we post-train and distill the video model into a single-step visual planner with fixed prediction horizon. Finally, we introduce action module via a mixture-of-transformers (MoT) architecture to learn implicit inverse dynamics model. Experiments demonstrate three key results: (1) VPP2-14B outperforms Cosmos3-64B by 11.0\% points in video prediction instruction-following success rate on open-ended tasks; (2) VPP2 surpasses the strongest baseline by 18.5\% points in success rate on real-world zero-shot ALOHA manipulation tasks; and (3) following benchmark-specific post-training, VPP2 achieves the highest success rates among evaluated methods on the challenging LIBERO-Pro, LIBERO-OOD, and RoboDojo benchmarks.

0.48

Explicit Geometric Chain-of-Thought for Vision-Language-Action in Autonomous Driving

cs.CV ·2026-10-07 ·Xingtai Gui, Yucheng Zhou, Dongqian Guo, Jiahao Gong, Feiyang Tan +1 more
code/data
Abstract · 21 pages, 9 figures. The code is available at https://github.com/TabGuigui/GeoCoTDrive

Vision-language-action~(VLA) models have emerged as a promising paradigm for autonomous driving. However, existing VLA models still suffer from a fundamental mismatch: driving actions require precise 3D geometric cues, while visual-language understanding and reasoning are largely conducted in a 2D semantic space. In this paper, we propose GeoCoTDrive, an explicit geometric chain-of-thought framework that grounds geometry in a planning-oriented manner. GeoCoTDrive follows a think with 2D first, drive with dedicated 3D priors paradigm. It first grounds 2D regions corresponding to decision-critical cues, and then retrieves localized 3D priors by sampling features from a geometric foundation model within the grounded regions. These localized geometric features are interleaved into the autoregressive context to support the trajectory generation. To supervise this process, we introduce planning-relevant grounding, a new region-level grounding task that focuses on local spatial cues directly affecting ego planning decisions, and construct the PlanningGrounding dataset to endow VLAs with planning-oriented grounding capability. Experiments across multiple end-to-end autonomous driving benchmarks show that GeoCoTDrive consistently improves safety-critical planning performance, demonstrating the effectiveness of the explicit geometric chain-of-thought process for VLA-based planning.

0.48

MESSENGER: Memory-Enhanced Sequential Scene Flow Estimation via Autoregressive Next-Frame Forecasting

cs.CV ·2026-10-07 ·Jiuming Liu, Jianing Li, Mengmeng Liu, Hongyang He, Hesheng Wang +1 more
top venuecode/data
Abstract · Accepted by NeurIPS 2026. Code will be released at: https://github.com/liujiuming123/Messenger

Scene flow can capture low-level 3D motion displacements in dynamic scenarios. Early pairwise estimators relying on instantaneous two-frame motion lack long-term temporal correlation and also struggle with poor extrapolation ability in future prediction. Although some recent methods attempt to explore multi-frame scene flow estimation in a sequence-to-sequence manner, they typically suffer from heavy computational overhead with increasing input frames and long-horizon prediction degradation due to ineffective motion propagation. To address these problems, we propose a novel memory-enhanced sequential scene flow pipeline, called MESSENGER. To sufficiently mine long-term temporal dependencies naturally within consecutive sequences, a memory buffer is designed by explicitly storing multiple history flow estimates and latent states. For each input frame, the temporally stored flows and states are correlated and retrieved to predict the current initialized flow in a next-frame forecasting manner. Furthermore, we develop an uncertainty-aware reweighting module to filter unreliable retrievals and mitigate accumulated errors. Extensive experiments on nuScenes and Argoverse 2 demonstrate state-of-the-art performance of our MESSENGER, reducing EPE3D by 71.6% on nuScenes and 67.7% on Argoverse 2 in long-horizon future extrapolation. This superiority can be attributed to our designed autoregressive forecasting paradigm, which naturally forces the network to progressively learn the next-frame distribution based on history observations. Code will be released at https://github.com/liujiuming123/Messenger.

0.47

Never Look Back: Understanding Persistence in 3D Object Memory from Egocentric Videos

cs.CV cs.AIcs.RO ·2026-10-07 ·Shravan Chaudhari, William Paul, Suchi Saria, Rama Chellappa, Homanga Bharadhwaj
code/data
Abstract · Project page: https://ledger-3d.github.io . Code: https://github.com/LEDGER-3D/LEDGER

As we move through the world and carry out everyday tasks, we encounter objects that may become relevant only later. We are capable of recalling where we left something or what was inside a container, even without knowing we would need it later. Here, we study how an embodied assistant can build a similar memory from egocentric videos, by observing a person's day-to-day activities. We present Ledger, a persistent 3D object memory that combines object locations, their histories, and contextual descriptions. It associates observations across the recording and retains objects after they leave the view, including those the person never touches. It clusters each object's observations by resting locations and records a move only after repeated evidence, reducing the effect of localization noise. Short descriptions preserve details such as an object's contents or supporting surface. It saves these records to later answer spatial questions without having to access the original images or video. Our memory raises HD-EPIC accuracy from 29.7% to 42.6%, UCS-Bench accuracy from 33.8% to 38.5% and localizes Ego4D objects with a 0.99 m median error on returned predictions. Our analyses identify complementary roles for temporal persistence, contextual descriptions, and retrieval. Our study on 100 stitched streams of multiple scenes each further exposes failures in both retrieval and construction. Per-scene construction partially recovers the performance lost across scene changes compared to that of single scene streams.

0.47

CoTrace: Data Recipes for Training Terminal Agents with Harness-Model Co-Evolution

cs.CL ·2026-10-07 ·Jixuan Chen, Jiaxin Zhang, Qinyuan Ye, Yada Pruksachatkun, Haoxiang Zhang +9 more
Abstract · Preprint. 32 pages, 7 figures, 17 tables

Terminal-agent capability depends jointly on model weights and the runtime harness that formats prompts, binds tools, and handles error recovery. Existing harness-model co-evolution approaches improve both components, yet often treat trajectories produced during harness search as an undifferentiated replay buffer. This practice overlooks that a trajectory's value for model training depends on the harness under which it was generated. To systematically analyze this interface, we establish an alternating co-evolution framework that decouples harness search and policy training through component-wise promotion decisions. Within this framework, we introduce CoTrace, a harness-aware data recipe that explicitly governs trajectory routing, provenance matching, and curriculum refresh. Under CoTrace, recurring execution failures guide harness synthesis, while policy training is strictly conditioned on verified rollouts matched to the adopted runtime for supervised fine-tuning (SFT) or fresh online interactions for reinforcement learning (RL). On the Tmax promotion split, CoTrace advances Qwen3.5-9B from 78 to 88 solved tasks under supervised fine-tuning while an online reinforcement variant reaches 90. Specifically, a compact harness-matched corpus produces steady model gains at substantially lower compute than much larger corpora pooled across sibling harnesses. Furthermore, evaluations on Terminal-Bench 2.1 and SWE-bench Lite show that out-of-distribution transfer depends fundamentally on harness compatibility, where maintaining consistency between training and evaluation runtimes prevents procedural execution breakdowns observed under foreign scaffolds.

0.46

Which Language Should a Skeleton Speak? Language Choices in Multilingual Reasoning

cs.CL cs.AI ·2026-10-07 ·HyeonSeok Lim, SeungWoo Song, Inho Won, Hoyun Song, Jihyo Kim +1 more
top venuebenchmarkcode/data
Abstract · Accepted to EMNLP 2026 (Findings)

Skeleton-based reasoning prompting is a promising training-free approach for structuring LLM reasoning, but prior work largely assumes an English-centric setting. We propose the Language-Aware Skeleton Exploration Framework (LASEF) to study skeleton-language choice in multilingual mathematical reasoning. Across math benchmarks, model scales, and languages, we show that English skeletons yield a small positive tendency on average, most visible for smaller models and low-resource languages. However, few language-level gains remain significant after correction, and English is not universally optimal. Combining greedy decoding, multi-rollout evaluation, translation ablation, and cross-benchmark validation, we further find three patterns of skeleton-language effects: directionally consistent, evaluation- and benchmark-dependent, and asymmetric negative. These effects cannot be fully explained by generation quality alone. Overall, skeleton language is a context-dependent design variable that requires multi-level exploration. All resources are released at https://github.com/lhsstn/LASEF.

0.46

Right Number, Wrong State? Measuring Cross-Jurisdiction Substitution in LLM Recall of State Policy

cs.CL ·2026-10-07 ·Jiayu Feng
unusual
Abstract · 6 pages, 3 figures, 1 table

When an LLM answers a state-specific policy question wrongly, it may be hallucinating, or it may be returning a real value that holds in another state. We test this with a minimal-set design: the question wording is fixed and only the jurisdiction varies, across the 50 U.S. states and the District of Columbia (51 jurisdictions) and three exactly defined Medicaid income-eligibility quantities. Gold values come from an official data book and agree with an independent source in 101 of 102 checked cells. Under a pre-registered protocol, Claude Sonnet 5.5 and GPT-5.6 Sol reproducibly give another state's current value, identical across two independent repeats, for 10 and 25 of 153 items. The count depends on the reference set: crediting any wrong answer that equals another state's value yields 3-5x more reproducible substitutions than checking every number in the asked state's own records, because many apparent cross-state answers are the asked state's own values under another convention or from an earlier year. Claims about cross-jurisdiction error need a complete same-state reference set. We will release the protocol, gold table, and all model outputs.

0.46

RLHND: Video Foundation Models as Physically Grounded Hand Trackers for Robot Learning

cs.CV cs.RO ·2026-10-07 ·Seungjun Moon, Subin Jeon, Sangwoo Kim, Hanbyul Joo, Jinwoo Shin
benchmark
Abstract · 34 pages, 15 figures

Recently, approaches that leverage human video datasets for robot policy training have become increasingly prevalent. However, most existing hand trackers regress pose from cropped frames with limited priors on hand motion and object interaction, resulting in inaccurate and physically inconsistent estimates. Moreover, the lack of physical cues, e.g., contact and force, limits the use of human videos for robot policy training. To this end, we propose RLHND, a video foundation model-based hand tracking model that jointly estimates hand pose and realistic tactile information from monocular egocentric videos. RLHND turns the pre-trained Cosmos 3 video diffusion backbone into a deterministic clip-level feature extractor via clean-latent conditioning, carrying its learned priors on hand motion and hand-object interaction into tracking. For pose estimation, RLHND (i) predicts hand poses with anatomically plausible joint angles and (ii) enables optional conditioning on the shape parameter to maintain consistent hand shape within the same video and even across videos recorded by the same actor. For tactile estimation, a separate tactile expert stream, trained with the pose stream frozen, predicts dense contact and force over the hand surface. We further adopt LBS-based feature spreading to enable vertex-wise feature extraction without costly per-vertex attention. RLHND achieves state-of-the-art performance across various benchmark datasets for pose estimation, while also achieving state-of-the-art performance in contact and force estimation. Moreover, we demonstrate the utility of RLHND for robot learning through retargeting results and real-world robot experiments. The code will be publicly available at https://seungjun-moon.github.io/rlhnd/.

0.46

Beyond Black Boxes: An Ontology Based Vision Framework for Transparent Fruit Disease Detection

cs.CV cs.AI ·2026-10-07 ·M. J. Marapperuma, B. V. K. I. Vidanage
top venue
Abstract

Efficiency and reliability in the timely identification of phytopathogens would help minimize crop loss in agricultural sectors but existing deep learning based methods, although accurate, would require a large annotated dataset and are black-boxed in their operation, thus, restricting flexibility and transparency. The current paper presents a hybrid detection framework that is able to overcome these drawbacks through the unified real-time vision with a semantic knowledge base. Web Ontology Language (OWL) was used to develop a comprehensive ontology of the disease of fruits that formally codifies phytopathological knowledge, such as symptom phenotype, spectral phenotype (HSV), and morphological phenotype. Such knowledge-guided configuration guides the image processing pipeline, making it possible to diagnose using this knowledge under expert specified rules without a large training dataset. The ontology underwent a computationally verified test of its structural soundness and commonsensical practical usefulness and scored an overall of 7.4/10. On the public set of MangoFruitDDS, the diagnostic performance was an F1 Score of 85.00%, precision of 85.00%, and a recall of 85.23%. These findings are very competitive and similar to other known deep learning models like the VGG16 that prove that the knowledge-driven approach provides an accurate, transparent, and scalable solution on automated fruit disease diagnosis.

0.45

TouchScale: 500 Hours of Human Vision and Touch for Visual-Tactile Learning

cs.CV ·2026-10-07 ·Dayou Li, Hao Wang, Qianqian Yang, Zihao Zhu, Haoquan Fang +21 more
code/data
Abstract · Project page: https://touch-scale.github.io/

Large-scale egocentric human interaction data is becoming an important source of physical supervision for embodied learning, yet video alone leaves the contact and pressure that characterize physical interaction unrecorded. Recent visual-tactile datasets provide this missing supervision, but their synchronized tactile data remain far smaller in volume than human video. Moreover, the largest resources often merge recordings from different sensors or annotation procedures, which makes the effect of data scale difficult to isolate. We therefore introduce TouchScale, a 500-hour dataset of contact-rich human interaction recorded with a single unified wearable setup. Its approximately 2K predefined task descriptions span everyday activities and structured manipulation, and each recording temporally aligns egocentric RGB-D video with wrist RGB video and dense full-hand bimanual tactile measurements. Compared with prior tactile data, training on the full TouchScale raises zero-shot contact IoU on data from an unseen tactile sensor from 0.134 to 0.383. Pretraining a visual encoder on TouchScale also yields the highest action recognition accuracy on three benchmarks among the compared visual-tactile datasets. Used for visual-tactile mid-training of a robot policy, TouchScale improves the average real-world success rate across four contact-rich manipulation tasks from 22.5% to 57.5%. With the sensor and collection protocol held fixed, both zero-shot tactile prediction and robot success show an overall upward trend as more TouchScale data is used. These results suggest that human visual-tactile data collected at scale with consistent sensing benefits both perception and robot manipulation. We will publicly release TouchScale, including all synchronized visual-tactile recordings and reconstructed object models, to support future research on scalable visual-tactile learning.

0.45

How Private is Private? A Comparative Study for Face De-Identification

cs.CV ·2026-10-07 ·Hui Wei, Hao Yu, Hui Kuurila-Zhang, Guoying Zhao
top venuebenchmarkcode/datahot direction
Abstract · Accepted to NeurIPS 2026. Project Page: https://cv-ac.github.io/hifd/

Face de-identification (FDeID) has emerged as a critical privacy-preserving technology, yet its evaluation remains fundamentally fragmented. Existing protocols rely on inconsistent metrics, heterogeneous datasets, and partial annotation coverage, so methods targeting different utility dimensions, such as landmark versus expression preservation, are reported on different benchmarks under different metrics, rendering cross-method comparison infeasible. We revisit FDeID evaluation from both the data and metric perspectives. On the data side, we introduce UtilFace, a curated, demographically balanced benchmark with high identity diversity, assembled from four large-scale face datasets through identity-aware cleaning, resolution enhancement, and stratified filtering. On the metric side, we propose HiFD, a Hierarchical Face De-identification metric that unifies identity suppression, multi-level utility preservation, and image quality under a single consistency-based paradigm: every component is computed from pretrained estimators' outputs on the original face and its de-identified counterpart, directly quantifying how much identity is suppressed and how much downstream-perceivable utility survives. HiFD organizes facial signals into a three-level utility hierarchy spanning macro cues (L1), micro cues (L2), and imperceptible cues (L3), and aggregates the five resulting components into a single interpretable score via weighted harmonic mean, with configurable application-specific profiles. Using this unified protocol, we conduct a comprehensive comparative study spanning adversarial, GAN-based, and diffusion-based methods, surfacing trade-offs and failure modes that remain invisible under existing protocols. We release the benchmark and evaluation toolkit to foster systematic and reproducible research in privacy-preserving human face analysis.

0.45

PhyDiCT: Plug-and-Play CT Reconstruction from Sparse X-Rays via Differentiable Rendering and Strong Priors

cs.CV ·2026-10-07 ·Weicheng Dai, Shantanu Ghosh, Kayhan Batmanghelich
top venuecode/data
Abstract · Accepted at MICCAI 2026; to appear in LNCS 16888

Reconstructing 3D Computed Tomography (CT) images from a few X-ray projections is a highly ill-posed inverse problem due to the loss of volumetric information. We propose PhyDiCT, a training-free framework that integrates a differentiable Physics-based forward model, grounded in the Beer-Lambert law, with a text-conditioned Diffusion as a strong prior to reconstruct 3D lung CT images. We refer to our approach as training-free since the prior model is used without fine-tuning, and our goal is to steer the denoising procedure to generate samples consistent with X-ray observations. We guide the diffusion generation using Split Gibbs sampling to jointly optimize for projection fidelity (reward) and consistency with prior knowledge. Also, we introduce a test-time refinement step that enhances image realism and anatomical coherence. We extensively evaluate our method on publicly available 3D CT datasets using both perceptual and semantic metrics, demonstrating that it surpasses existing plug-and-play diffusion and fully trained reconstruction approaches. Our findings highlight that combining a strong generative prior with the underlying physics of image formation substantially improves reconstruction quality, e.g., 7.5\% improvement on SSIM compared to full training methods. Code will be released at https://github.com/batmanlab/PhyDiCT.

0.44

SPLIT-RL: Staged Perception-Language Reasoning Training with Claim-Level Advantages

cs.CV ·2026-10-07 ·Raja Kumar, Rajat Koner, Ritwick Chaudhry, Zhuowei Li, Nishant Sankaran +1 more
Abstract

Vision-Language (VL) reasoning requires a model to both extract relevant and accurate information from an image (visual reasoning, VR), and to infer the answer from it (language reasoning, LR). Reinforcement learning with verifiable rewards typically trains both through a single chain-of-thought with a final-answer reward. This gives every CoT token the same sequence-level advantage, failing to distinguish capability specific errors. We propose SPLIT-RL, a staged post-training approach that trains VR and LR in disjoint phases. Because a group's rollouts differ along one capability at a time, the group-relative advantage isolates it, and each phase is optimized using phase-specific reward. We further introduce Claim-Level Advantage (CLA-GRPO), which decomposes VR-phase rollouts into atomic visual claims and provides a fine-grained advantage at claim level based on visual-type group formation. Although trained in two phases, trained policy is evaluated like GRPO model, with a single CoT call at inference time. Under this protocol, SPLIT-RL improves average accuracy over GRPO by 1.4-6.1 points across Qwen3-VL models from 2B to 30B-A3B and InternVL3.5-8B. Evaluating each capability using an oracle based diagnostic shows that answer-only GRPO leaves perception unchanged, whereas SPLIT-RL improves both VR and LR.

0.44

Entity-Aware Evaluation of OCR Systems for Historical Document Transcription

cs.CL cs.AIcs.CVcs.LG+1 ·2026-10-07 ·Camilo Mariño, Diego Belzarena, Marina Gardella, Ignacio Ramírez, Gregory Randall +1 more
publishedbenchmark
Abstract · Accepted at URUCON 2026

Optical Character Recognition (OCR) is essential for digitizing historical archives, yet standard evaluation metrics such as Character Error Rate (CER) do not necessarily reflect whether meaningful information, such as names, dates, and identifiers, is preserved. In this work, we investigate entitylevel analysis of OCR errors as a complementary perspective for evaluating OCR quality in historical documents. Using a publicly available benchmark of 175 documents from the Berrutti Archive, augmented with 5,209 manually annotated entity mentions that will be released upon acceptance, we compare sequence-recognition OCR systems, OCR-specialized vision-language models, and general-purpose vision-language models. Beyond global transcription accuracy, we analyze entity preservation and recoverability under varying character-level edit tolerances. Our experiments show that OCR rankings depend on the evaluation criterion: while Qwen 3.6 achieves the highest entity coverage across all tested tolerances, several systems with lower exact-match coverage recover substantially when small errors are tolerated, suggesting that many of their errors correspond to localized character-level deviations rather than fundamentally incorrect entity transcriptions. Conversely, some systems remain unable to recover many entities even under relaxed matching. These findings show that OCR quality cannot be fully characterized by aggregate transcription metrics alone and highlight the importance of evaluating how OCR errors affect the preservation and recoverability of meaningful information.

0.43

LVSPM: Long Sequence View Synthesis and Pose Estimation Model

cs.CV ·2026-10-07 ·Xi Chen, Yachi Zhang, Linghao Chen, Minghua Liu, Hao Su +2 more
top venuecode/data
Abstract · ECCV 2026. Project Page: https://burningdust21.github.io/Projects/LVSPM/

We present LVSPM, a generalizable model that jointly estimates camera poses and synthesizes novel views from uncalibrated image collections. Trained with only RGB images and pose supervision, LVSPM avoids dense 3D ground truth and employs test-time training (TTT) layers to scale seamlessly to hundreds of input views. On RealEstate10k, Co3Dv2, and DL3DV, LVSPM surpasses VGGT in pose estimation across 16-256 views, with especially large margins at strict thresholds. For novel view synthesis under a practical protocol where more views cover larger scenes, LVSPM achieves state-of-the-art pose-free quality---surpassing even pose-dependent models in PSNR---and still maintains high quality as scene scale grows, while baselines collapse. The code is available at https://burningdust21.github.io/Projects/LVSPM .

0.43

ORCA: Hunting Compositional Failures in Text-to-Image Diffusion

cs.CV cs.LG ·2026-10-07 ·Arshia Hemmat, Amirhossein Vahidi, Amitis Shidani, Mohammad Vali Sanian, Hesam Asadollahzadeh +2 more
top venuepositiontheorycode/data
Abstract · Accepted at NeurIPS 2026. Camera-ready version. 28 pages, 4 figures, 13 tables. Project page: https://arshiahemmat.com/orca

Text-to-image diffusion models fail predictably on compositional prompts: attributes bind to the wrong objects, spatial relations invert, and multi-object scenes lose count. Recent architectures already augment CLIP with a T5 encoder precisely because CLIP's contrastive embedding loses compositional structure, yet these failures persist. We argue the binding problem is therefore not one of missing information but of misaligned information: a text encoder preserves compositional structure, but in a representation space shaped by language modelling rather than vision, and the denoising objective does not directly reward aligning the two. We show this correspondence can be supplied as an explicit training signal, that the relevant cross-modal information is concentrated in a low-rank subspace of self-supervised visual features, and that supplying it can be folded into diffusion training as a single auxiliary loss. Our method, ORCA (Orthogonal Residual Compositional Alignment), aligns the latent of a diffusion transformer with a low-rank target derived from a frozen visual encoder, through a predictor whose orthogonal basis is parameterised by a learned residual between T5 and CLIP embeddings, which provides a prompt-dependent signal for selecting the visual readout subspace. We prove that the cross-modal information recoverable at a given rank is bounded by the spectral mass of the visual encoder's covariance in the top components. Across three diffusion-transformer backbones (DiT-B/2, DiT-L/2, U-ViT-L), ORCA improves FID and GenEval over both vanilla and REPA baselines at zero inference-time cost; on DiT-L/2 it reaches FID 16.65 and GenEval 0.291 at 200K steps, exceeding the strongest 400K baseline at half the training cost, with the largest gains concentrated on attribute binding, spatial relations, and multi-object prompts.

0.42

A Comparative Study of Evaluation Metrics for Long-Document Financial Narrative Summarization with Transformers

cs.CL ·2026-10-07 ·Nadhem Zmandar, Mo El-Haj, Paul Rayson
publishedbenchmarkhighly cited
Abstract · 12 pages

There are more than 2,000 listed companies on the UK's London Stock Exchange, divided into 11 sectors who are required to communicate their financial results at least twice in a single financial year. UK annual reports are very lengthy documents with around 80 pages on average. In this study, we aim to benchmark a variety of summarisation methods on a set of different pre-trained transformers with different extraction techniques. In addition, we considered multiple evaluation metrics in order to investigate their differing behaviour and applicability on a dataset from the Financial Narrative Summarisation (FNS 2020) shared task, which is composed of annual reports published by firms listed on the London Stock Exchange and their corresponding summaries. We hypothesise that some evaluation metrics do not reflect true summarisation ability and propose a novel BRUGEscore metric, as the harmonic mean of ROUGE-2 and BERTscore. Finally, we perform a statistical significance test on our results to verify whether they are statistically robust, alongside an adversarial analysis task with three different corruption methods.

0.42

VolCo: Volumetric Contact for High-Fidelity Human Grasp Generation

cs.CV ·2026-10-07 ·Zhuo Chen, Yihua Cheng, Aleš Leonardis, Hyung Jin Chang
top venuebenchmarkcode/data
Abstract · Accepted to NeurIPS 2026

Accurate contact modeling is fundamental to understanding hand-object interaction, yet existing contact representations are typically restricted to object surfaces and rely on hand-crafted rules to recover contact details, leading to severe penetrations and implausible results. To better exploit the rich detail in motion-capture data, we introduce Volumetric Contact (VolCo), a representation that expands surface points to a set of 3D volumetric grids. VolCo encodes 3D contact that allows precise hand part recovery, and is organized in an inherent hierarchy: local contact details within each volume and global hand geometry across all volumes. Our framework, VolCoDiff, employs two modules to capture local and global features following this hierarchy. For local contact details, we use a 3D variational autoencoder to model the possible hand configurations conditioned on the local object signed distance field (SDF). For global hand geometry, we design a prior-guided diffusion model that learns the distribution of compressed latent features aggregated from the volumetric grids. We evaluate our method on two benchmark datasets and demonstrate state-of-the-art performance in penetration and stability, indicating the capability to generate tight grasps with much less severe penetrations. Our code is available at https://github.com/chzh9311/volco.

0.42

EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory

cs.CL ·2026-10-07 ·Hongru Cai, Ran Wei, Wenjie Wang, Chengfa Wu, Ning Song +2 more
Abstract

Conditional memory architectures such as DeepSeek Engram use input n-grams to look up learned embeddings, expanding the capacity of large language models (LLMs) with limited additional computation. Beyond model scaling, this architecture has demonstrated the potential to decouple factual knowledge storage from general-purpose computation, offering a promising route to updating factual knowledge while keeping the Transformer backbone fixed. Realizing this potential is challenging because different expressions of a fact may activate different n-gram embeddings, while updating shared embeddings can unintentionally change the model's predictions about other facts. We propose EngramEdit for decoupled knowledge updates through conditional memory. EngramEdit first computes target memory representations that make the model predict the updated fact across multiple expressions. It then jointly updates the shared n-gram embeddings to match these targets across expressions and edits, penalizing updates to frequently reused embeddings more strongly to preserve unrelated knowledge. Experiments show that EngramEdit enables independent factual knowledge updates through conditional memory, achieving near-perfect editing success. Revised knowledge is usable across unseen expressions and in multi-hop reasoning, with nearly three times the strongest baseline's accuracy under chain-of-thought (CoT) prompting. Unrelated knowledge and general capabilities are largely preserved even as factual updates accumulate. These findings show that EngramEdit turns conditional memory into an editable knowledge interface, extending its role beyond model scaling to support decoupled knowledge updates.

0.41

The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models

cs.CL ·2026-10-07 ·Obada Kraishan
publishedbenchmarkcode/data
Abstract · Accepted at the 2026 IEEE 8th International Conference on Cognitive Machine Intelligence (IEEE CogMI 2026). 8 pages, 3 figures, 5 tables. Co…

Reasoning models allocate extra computation at inference time and present their answers as the product of deliberate thought. If this deliberation works the way dual-process accounts of human cognition suggest, longer thinking should weaken the classic decision biases that fast, intuitive judgment produces. Using 30 vignettes covering six biases (anchoring, framing, loss aversion, escalation of commitment, availability, confirmation) from an established benchmark, we run a dose-response study across four model families, pairing each reasoning model with a matched non-reasoning sibling and requesting thinking ceilings of 0, 1,024, 4,096, and 8,192 tokens, for 12,350 API calls. Because a requested ceiling is not the same as realized deliberation, we use the reasoning tokens each call consumed as the dose. First, reasoning models are not less biased than their siblings; the point estimate leans the other way in every family, but the item-level pooled contrast is not reliable (Delta = +0.031, t(29) = 1.45, p = .157). Second, bias magnitude does not reliably fall as realized deliberation grows: no slope is significantly negative, and where anything moves it is the signed score drifting further from the human direction. Third, anchoring is the only bias in the human direction (d = 1.89). Four of the other five lean the opposite way in all seven models; with five items per bias, that reversal is reliable for framing and directional for escalation of commitment, confirmation, and loss aversion, while availability is absent. A one-line instruction to restate the anchor before answering lowered anchoring on all five anchoring items, which no amount of additional thinking did, although the effect does not reach significance (p = .057). The results argue against treating test-time reasoning as a rationality guarantee and for auditing deployed models bias by bias.

0.41

Goldsmith: Gold-Loss-Guided Definition Optimization with an Agentic Annotation Harness

cs.CL ·2026-10-07 ·Yihan Li, Hanyi Zhang, Xiaoxi Jiang, Man Guo
top venue
Abstract · 20 pages, 4 figures, 11 tables. Accepted to the main conference of EMNLP 2026

Many annotation projects begin before experts have a stable guideline or enough labels to train a task-specific model. We present Goldsmith, an agentic pipeline that turns a small gold set---expert-annotated calibration examples representing the intended task boundaries---into a reusable structured annotation definition. Goldsmith treats this definition as a trainable textual object. Candidate definitions are run on the same gold examples and scored with an executable structured loss, while the output schema, formatting, retrieval, repair, judging, and human review remain in an external harness. A large language model (LLM) editor converts the highest-loss failures into textual-gradient revisions, which are accepted only when the measured loss decreases. In prompt-optimization comparisons, Goldsmith improves over direct rewriting, OPRO, APE, and PromptBreeder under matched evaluation protocols. The resulting definition also improves downstream annotation when combined with retrieval, score-based routing, and human review across typed span, pair-level relation, and fixed-trigger event-argument tasks. These results show that scarce expert supervision can support both task-definition learning and scalable annotation.

0.40

MOTIF: Person-of-Interest Deepfake Detection Beyond 3DMM Coefficients

cs.CV ·2026-10-07 ·Giovanni Affatato, Sara Mandelli, Paolo Bestagini, Stefano Tubaro
benchmarkcode/data
Abstract · 6 pages. Accepted at the 2026 IEEE International Workshop on Information Forensics and Security (WIFS)

Video deepfakes targeting a specific individual, the Person-of-Interest (POI), are the most harmful ones, and, since a public figure is abundantly recorded, a detector can be built from genuine footage of that individual. Such detectors commonly describe a subject through a 3D Morphable Model (3DMM) and adopt its coefficients as a whole, so which part of that description carries the signal has never been measured. We dissect it, holding the encoder, the training corpus and the enrollment protocol fixed and varying only what the encoder observes. The groups of coefficients prove largely redundant, since the shape block alone recovers almost all the accuracy of the full vector, and their temporal evolution contributes a real but bounded amount. We further show that the dense surface the same fit returns, which these detectors discard, carries identity information that the coefficients do not, and that it helps precisely where they are weakest. We assemble the best configuration into MOTIF, a visual-only detector trained on real videos only, with no manipulated video and no POI-specific data. It improves on both state-of-the-art POI detectors in every dataset and manipulation of our benchmark and at two quality levels. Our experimental code will be released at https://github.com/polimi-ispl/MOTIF.

0.40

VIS-Ground: Video Interactive Storytelling with Contextual Grounding

cs.CV ·2026-10-07 ·Bingxuan Li, Yiwen Song, Xueqing Wu, Yanzhou Pan, Yang Li +8 more
code/datacommunity pick
Abstract · Project Page: https://vis-ground.github.io

Video interactive storytelling enables viewers to actively steer how a video unfolds. However, once we allow viewers to intervene during generation, a new challenge arises: The viewer's request can have latent dependencies on both the grounding source and the current rendered video state. These dependencies may not be explicitly stated in any individual input, but emerge only when the source, rendered history, and new viewer intent are considered jointly. Existing interactive video generation systems primarily emphasize following viewer instructions, while source-grounded video generation methods focus on aligning generated content with an external narrative or knowledge source. This leaves a fundamental question underexplored: What context should a generation model ground on during interactive continuation, and how can heterogeneous, unstructured inputs be transformed into such grounding context? In this work, we formulate contextual grounding as the process of transforming heterogeneous input context into an executable constraint model for video generation. To address this challenge, we introduce VIS-Ground, which performs Structured Context Abstraction to recover grounded states and cross-context dependencies, Generation Constraints Induction to project relevant dependencies into candidate-specific constraints, and Constrained Video Generation to enforce these constraints through planning, verification, revision, and rendering. Across three video generation backbones, VIS-Ground consistently achieves the highest overall composite score, reaching an average absolute improvement of 10.3 points over the strongest per-backbone baselines. Detailed analysis further shows gains across both narrative and knowledge grounding, and reveals remaining challenges in dependency extraction, and faithful realization during video rendering.

0.40

A Multi-Source Ultrasound Benchmark Revealing the Limits of Contemporary Self-Supervised Anomaly Detection Methods

cs.CV cs.LG ·2026-10-07 ·Marco Riedenauer, Daniel Kienzle, Pratik Mayekar, Rainer Lienhart
publishedbenchmark
Abstract · 7 pages, 3 figures

Self-supervised anomaly detection is a promising paradigm for medical ultrasound, as normal images are often easier to obtain than exhaustive annotations of all possible pathologies. However, most existing evaluations are limited to a single anatomy or task, making it unclear whether models learn a robust notion of normal ultrasound appearance or only a source-specific representation. We introduce the SADUSI benchmark, a multi-source ultrasound dataset designed to train and evaluate anomaly detection methods across a broad range of anatomical regions, views, and acquisition protocols. The goal of SADUSI is to provide a diverse normal ultrasound distribution and a benchmark for visible structural anomalies that can be assessed from single images. We evaluate representative self-supervised anomaly detection methods and find that current approaches struggle in this setting. In particular, reconstruction-based diffusion methods such as AnoDDPM and DeCo-Diff achieve pixel-level AUROC values of 0.56-0.72 and maximum F1 scores of 0.10-0.26, indicating limited separation of pathology from normal image regions. Feature-based PatchCore variants perform better, reaching pixel-level AUROC values of 0.76-0.83, but remain limited with maximum F1 scores of 0.14-0.40. These findings suggest that broad multi-source ultrasound anomaly detection remains an open challenge and that SADUSI can serve as a resource for developing methods that generalize beyond anatomy-specific settings.

0.40

GPU-Accelerated Computation of Persistent Homology for Topological Analysis of Image Data

cs.CV ·2026-10-07 ·Fan Wang, Hubert Wagner, Rezaul Chowdhury, Chao Chen
top venuebenchmarkcode/data
Abstract · Accepted for publication in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). Includes supplementary material (8 pages…

In recent years, persistent homology has seen rapid adoption in deep learning, yet its computation remains a major bottleneck in network training. This paper introduces TopoGPU, a GPU streaming pipeline that computes persistence diagrams of cubical complexes induced by 2D and 3D images. TopoGPU streams the input image chunk by chunk, processing each chunk with massively parallel GPU kernels on a grid of GPU blocks; the resulting boundary relations are accumulated in host memory, where the CPU performs the boundary matrix reduction. TopoGPU introduces a stratification-aware discrete Morse matching that provably preserves persistent homology under streaming, together with a parallel topological sorting algorithm and a parallel V-path parity algorithm for deriving Morse boundaries on the GPU. TopoGPU outperforms Cubical Ripser, a state-of-the-art method for persistent homology computation, on every benchmark evaluated, achieving an average end-to-end speedup of 53.24x and a maximum of 198.01x. We further integrate TopoGPU into a topology-preserving deep network, demonstrating that it substantially reduces the cost of persistent homology computation during network training. TopoGPU is open source, with pre-built binaries, Google Colab notebooks, and Docker images available at the project's GitHub page: https://github.com/seravee08/GPU-Computation-of-Persistent-Homology-for-Image-Data.

0.40

Shaer: Controlled Arabic Poetry Generation with Meter Subform and Semantic Conditioning

cs.CL ·2026-10-07 ·Ahmad Abbas, Tamara Fakih, Nour Fakih, Ammar Mohanna
publishedcode/data
Abstract · 22 pages, 7 figures. Code: https://github.com/AhmaddAbbass/Shaer ; models and datasets: https://huggingface.co/Shaer-AI

Classical Arabic poetry generation requires simultaneously satisfying semantic, linguistic, and fine-grained prosodic constraints. Existing systems typically control broad poetic attributes but do not jointly model semantic intent, meter subform, and poem length. We present Shaer, a controllable Classical Arabic poetry generation framework jointly conditioned on natural-language descriptions, meter subforms, and target hemistich counts. To support this task, we construct an enriched corpus of 116,032 classical Arabic poems derived from Ashaar, containing normalized meter-subform labels and automatically generated, validated semantic descriptions. We then adapt Yehia-7B using QLoRA-based supervised fine-tuning with a completion-only objective. Our evaluation combines automatic assessment of base-meter conformity, requested-subform adherence, and length control with three LLM judges, blinded human evaluation, and memorization analysis. Shaer achieves 95.17% base-meter accuracy, 91.75% poem-level meter-subform accuracy, and 83.40% exact count accuracy. Relative to its untuned foundation model, these results represent gains of 68.68, 57.77, and 38.93 percentage points, respectively; Shaer also attains the highest base-meter accuracy among all evaluated systems. Multi-LLM evaluation and a blinded human assessment of top-ranked outputs further indicate competitive semantic and literary quality. Finally, analysis of all 3,481 test generations finds no exact copies from the training corpus or paired source poems. Code, models, and datasets are publicly available.

0.40

DOGS: Design-Space Sampling for Prompt-Driven Logo Generation

cs.CV ·2026-10-07 ·Ganyu Zou, Chen Dai, Nathan Self, Kevin Piper, Ramachandra Rao Seethiraju +3 more
publishedpositioncode/data
Abstract · Accepted to BMVC 2026

Prompt optimization for text-to-image (T2I) generation has been pursued almost entirely as text rewriting, in which a short user brief is expanded into a longer, model-preferred token sequence. We argue that such a language-space formulation is ill-suited to structured visual design tasks such as logo creation, where a one-line brief leaves most design decisions unspecified. These decisions depend on relational priors that a linear sequence cannot encode, and they leave an uncontrolled channel through which protected marks may be reproduced. We therefore recast logo prompting as sampling within a structured design space, and instantiate this idea as DOGS (Design-space prompting with an Originality-aware GFlowNet Sampler). From a large corpus of real-world logos, we mine a typed, graph-structured design grammar whose edges record empirical co-occurrence. A GFlowNet sampler then generates design graphs with probability proportional to a terminal reward that combines recognizability, aesthetics, and corpus-relative originality. Every slot draws only from a closed design-level vocabulary, and any infringement-inducing or harmful token is removed during parsing. The originality reward further penalizes proximity to existing logos, thereby incorporating infringement avoidance into the method by construction. On two open-source renderers and against nine baselines, DOGS produces logos that are more recognizable and aesthetic, substantially more diverse, and far less prone to trademark infringement.

0.40

Gaussian Density Splatting Network

cs.CV ·2026-10-07 ·Miao Shang, Yabin Wang, Xiaopeng Hong
code/dataunusual
Abstract · This is the preprint version of the paper and supplemental material to appear in NeurIPS, 2026. Please cite the final published version

This paper proposes a novel crowd counting approach, the Gaussian Density Splatting Network (GDSNet). Unlike methods that rely on conventional, grid-based density maps and are sensitive to spatial resolution, GDSNet represents a crowd as a superposition of continuous 2D Gaussian primitives. Our approach is built upon two key contributions. First, a control-point-based fitting mechanism is introduced to structure the prediction of Gaussian parameters. A set of control points is adaptively allocated to define local regions, from which features are pooled to regress each primitive's parameters. Second, we adapt a differentiable Gaussian Splatting framework to the counting task by parameterizing each primitive with geometric parameters and a scalar density mass. This formulation allows the network to be trained end-to-end via spatial matching of differentiably rendered density maps, naturally providing both local density supervision and global count optimization. Extensive evaluations on four standard benchmarks show GDSNet consistently outperforms the state of the art. The code is available at https://github.com/infinite0522/GDSNet-Gaussian-Density-Splatting-Network.

0.39

Unbalanced Optimal Transport for Efficient Visual Document Retrieval

cs.CV ·2026-10-07 ·Hoyeon Shin, Jeongyeon Kim, Yeong Jun Koh, Yeoneung Kim, Hanul Kim
top venuecode/data
Abstract · Accepted at ECCV 2026

Vision-Language Models(VLMs)have advanced Visual Document Retrieval (VDR) by enabling OCR-free, multi-vector representations. However, these models generate excessive tokens to preserve dense visual details, which severely limits retrieval scalability. Existing compression methods mitigate this overhead but rely on isolated heuristics that fail to balance token importance, semantic redundancy, and coverage capacity. We propose a visual token compression framework formulated as an Unbalanced Optimal Transport (UOT) problem. Our approach unifies these three objectives into a single alternating optimization process. Furthermore, we theoretically show that minimizing the transport cost upper-bounds the MaxSim retrieval error, ensuring ranking stability. Extensive experiments on the ViDoRe V1 and V3 benchmarks demonstrate that our method preserves fine-grained document semantics and improves the efficiency-effectiveness trade-off over state-of-the-art base-lines. Code is available at https://github.com/shhhhhyy/Unbalanced-Optimal-Transport-for-EVDR.

0.39

Closing the Loop on Contrail Avoidance with Satellite Verification

cs.CV cs.LG ·2026-10-07 ·Spandan Ghose Chowdhury
publishedunusual
Abstract · Accepted into Tackling Climate Change with Machine Learning: workshop at NeurIPS 2026

Contrails are the thin ice clouds that aircraft leave behind. They cause a large share of aviation's warming, and rerouting the few flights that produce them could avoid much of it. However, an avoided contrail only counts if a satellite can confirm that it never formed, and this check is hard: contrails are one to two pixels wide, cover only 0.18% of pixels, and look very similar to natural cirrus. We build a small diffusion model (8.4M parameters, trained on one GPU) that detects them, and we run a controlled study to find out which components matter. The model reaches 0.476 PR-AUC, compared with 0.414 for a DeepLabV3+ baseline and 0.119 for an adapted MedSegDiff. Doubling the input resolution of the CNN brings it to parity (0.499, p=0.07). Three lessons apply beyond contrails. First, check the input resolution before designing a new architecture. Second, simple flips and rotations more than double accuracy and matter more than any architectural choice we measured. Third, pretraining the model on contrail shapes is harmful: the model learns that thin strokes appear everywhere and paints them onto empty scenes. Precision collapses to 1% while recall-based metrics still rate the degraded model as excellent, and no threshold or guidance heuristic repairs this failure.

0.38

Toward Part-Aware Choral Transcription with singing voice assignment

eess.AS cs.SD ·2026-10-07 ·Hanyu Meng, Zhanhong He, Zixun Guo, Yaolong Ju
code/data
Abstract · Submitted to ICASSP2027

Single-instrument automatic music transcription (AMT) has advanced substantially, yet choral applications require soprano, alto, tenor, and bass (SATB) to be transcribed as separate parts. Recent note-level choral AMT instead produces a single merged note track, limiting rehearsal, education, and score reconstruction. To address this limitation, we introduce Part-aware Choral Transcription (PawCT), to our knowledge the first end-to-end neural framework that identifies active SATB parts from choral audio and transcribes each into a separate note-level track. PawCT combines part-specific onset, offset, and frame prediction with part-presence estimation, union-level supervision, and structured training targets using a range prior (RP) based on SATB pitch ranges and its ordered-continuity (OC) extension, which adds within-part melodic continuity and cross-part pitch ordering. On YouChorale, PawCT-RP-OC achieves a macro part-aware note F1 of 0.225 at a 50-ms onset tolerance, outperforming an adapted choral baseline (0.165) by 36.4% relative and a two-stage post-hoc assignment pipeline (0.175). Its part-agnostic variant, PagCT, achieves a 50-ms onset F1 of 0.382, compared with 0.237 for the previous state-of-the-art choral AMT model. Cross-dataset evaluations on CSD and Cantoria further assess performance under dataset shift. These results demonstrate the benefit of jointly modeling note transcription and vocal-part assignment. Code and demos are available at https://hanyu-meng.github.io/Paw_Choral_AMT_Demo/.

0.37

TERRA: Learning Transportable Latent Actions through Temporal Effect Representation and Relational Alignment

cs.CV cs.LGcs.RO ·2026-10-07 ·Tianxingjian Ding, Mubarak Shah, Yu Tian
positioncode/data
Abstract · Preprint. Code and project page coming soon

Latent actions supervise robot policies with action-like codes inferred from visual transitions, and their usefulness hinges on two questions: what a code keeps from a transition, and whether it still means the same thing when reused in a different initial state. The first is a tension in time: an endpoint difference discards how motion unfolds, while the full sequence admits nuisance variation. The second is left open by reconstruction, which only ever observes a latent together with the state it came from. We argue that both questions can be answered in the same place. TERRA (Temporal Effect Representation and Relational Alignment) describes a transition by a compact temporal effect, its net feature change together with a low-order within-window dynamics component, and learns a continuous latent from this effect. The same effect space then serves as the reference for reuse: Effect-Anchored Transport (EAT) decodes a latent in other initial states and anchors the resulting effect to the one observed at its source, so that the latent is shaped by what it does across contexts rather than only by the transition it came from. With frozen linear readers, TERRA predicts actions more accurately than UniVLA and a LAPA-style baseline, degrades more slowly under visual distractors, and keeps transported transitions faithful to the donor action as the recipient context moves farther away; a same-budget control shows that these gains come largely from EAT. At matched pretraining scale, the complete system reaches 93.4% average success on LIBERO, compared with 91.8% for UniVLA.

0.37

VICO: Visual Environments Co-Evolving for Vision-Language Model Reasoning

cs.CV cs.AI ·2026-10-07 ·Meng Lu, Ligeng Zhu, Olivia Xiao, Yuchen Zhuang, Zihan Wang +6 more
position
Abstract

Reinforcement learning with verifiable rewards (RLVR) has become a standard recipe for post-training vision-language models (VLMs), but it typically assumes a static training environment. As the actor improves, fixed tasks drift out of its learning frontier: many become trivial, others remain unsolvable; and the learning signal collapses. We argue that VLM post-training should evolve the visual environment alongside the actor, not just the actor itself. We propose VICO, a co-evolutionary framework in which an actor and an Environment-as-Rewriter (EnvRewriter) are trained jointly: the EnvRewriter edits verifiable image-side structures, such as scene graphs, chart tables, or protected region masks, and re-renders them to produce label-valid training samples whose difficulty is calibrated to the actor's current ability through a pass-rate-based reward. This loop continuously realigns task difficulty with actor capability without any additional human annotation. Across nine multimodal benchmarks spanning mathematical reasoning and visually grounded understanding, VICO-8B improves over its base model by up to +5.0% on out-of-domain tasks, surpasses the strongest self-evolution and text-editing co-evolution baselines by +4.3% and +8.4% respectively, and stays comparable to chart-specialized RLVR methods using 16-160 times fewer labeled samples. By shifting from human-labeled supervision to image-editing co-evolution, VICO offers a scalable path beyond static-corpus RLVR for visual reasoning.

0.37

Quasi-Binarized Autoencoders: An Architecture-Independent Information Bottleneck for Medical Image Anomaly Detection

cs.CV cs.LG ·2026-10-07 ·Shouhei Hanaoka, Takahiro Nakao, Atsushi Takamatsu, Takeharu Yoshikawa, Osamu Abe
benchmarkcode/data
Abstract

Unsupervised anomaly detection, which learns only from normal images, is a central task in medical image analysis and remains an open problem. Reconstruction-based methods pass an image through an encoder-decoder network trained on normal data and detect anomalies from the residual between the image and its reconstruction. This works only if the information passed from the encoder to the decoder is limited; otherwise the network learns an identity mapping and reconstructs anomalies too. This limit is usually imposed through architectural choices, tuned per dataset, that cannot be stated in bits. We introduce the quasi-binarizing (QB) layer, which squashes each latent element into [0, 1] and adds Laplace noise of scale 1/epsilon. Each element is then epsilon-locally differentially private, and the mutual information between an image and its reconstruction is bounded by a quantity that depends only on epsilon and the number of QB elements, whatever the encoder and decoder. Placing a QB layer on every encoder-decoder path, including all skip connections, we build QBAE, a seven-level attention U-Net with 32,768 QB elements. On the seven datasets of the MedIAnomaly benchmark, QBAE with one architecture and one configuration reaches a mean image-level AUROC of 0.828, the highest among methods that do not adapt to each dataset, and the best reported results on BraTS2021 (AUROC 0.911, pixel-level AP 0.838). The noise is kept at test time, so that every reconstruction satisfies the bound. Without input corruption, the bottleneck alone prevents identity collapse (mean AUROC 0.805 vs. 0.590). Code is available at https://github.com/hanaokalog/MedIAnomalyQB.

0.37

Unsupervised Maneuver-Aware Acoustic Fault Detection for Autonomous Drones

cs.SD eess.SY ·2026-10-07 ·Ali M Ali, Ziyi Tang, Nurdaulet Nazarbay, Hashim A. Hashim
top venue
Abstract

This paper presents a maneuver-aware acoustic fault detection framework for autonomous drones that integrates Noise2Noise-inspired deep learning denoising with maneuver-conditioned reconstruction. A key practical constraint motivating this work is that labeled faulty-flight data are difficult and potentially unsafe to collect; the proposed framework therefore follows an unsupervised learning paradigm in which only nominal flight recordings are required for training. In flight environments, acoustic signals acquired from unmanned aerial vehicles are subject to variability arising both from environmental noise and from structured, maneuver-dependent aerodynamic effects. To address these challenges simultaneously, a two-stage learning architecture is developed. In the first stage, a Noise2Noise-inspired denoising model attenuates stochastic acoustic noise while preserving fault-relevant spectral-temporal structures, without requiring clean reference signals. In the second stage, a maneuver-Conditioned Convolutional AutoEncoder (maneuver-CCAE) is trained using maneuver-related labels including drone type and flight direction to model nominal acoustic behavior under varying operating conditions. Fault detection is subsequently performed using reconstruction error as an anomaly score. Experimental results demonstrate that the proposed maneuver-aware conditioning raises the area under the ROC curve (AUC) from $\AUCaeOnly$ (unconditioned baseline) to $\AUCfull$ (full model), validating the critical role of maneuver-dependent modeling. The complete framework is deployed on an NVIDIA Jetson Orin Nano Super embedded platform within a ROS2 pipeline, achieving an end-to-end fault detection latency of approximately $20\,\text{ms}$ per audio segment with a TensorRT half-precision (FP16) backend, confirming real-time viability for onboard UAV health monitoring.

0.37

Does Document Structure Help Dense Retrieval? A Placebo-Controlled Ablation of Four Mechanisms Across Two Corpora

cs.IR cs.AI ·2026-10-07 ·Andrey Kuehlkamp, Priscila Correa Saboia Moreira, Samuel Rund
Abstract · Initial draft,

Retrieval-augmented generation systems increasingly rely on document-structure treatments: structure-aligned chunking, LLM-generated chunk contexts, heading-path metadata, and hierarchical two-stage retrieval. Separate studies support each on different corpora, embedders, and metrics, and none control for a shared confound: any text prepended to a chunk perturbs its embedding. We present a mechanism-isolating ablation testing all four treatments under one protocol, matching chunk sizes across conditions and adding a semantically null placebo---heading paths that are structurally valid but shuffled across documents. We score retrieval with a coverage-aware nDCG and test four pre-registered contrasts via document-clustered bootstrap with Holm correction, on two distant corpora: 200 Wikipedia Featured Articles (951 queries) and 1,585 QASPER papers (4,303 questions). Organization helps, and the cause is content, not tokens: structure-aligned chunks with real heading paths beat contextualized fixed windows (+0.022 / +0.012 cov-nDCG@10) and the placebo (+0.010 / +0.016). Naive two-stage hierarchical retrieval hurts (-0.033 / -0.015), traceable to first-stage section recall. Gold structure beats LLM-induced structure on Wikipedia but not on QASPER. Effects are small ($dz$ 0.06-0.11) but Holm-significant and consistent across corpora.

0.36

InsClaimBench: Benchmarking Insurance Claim Adjudication Across the Decision Chain

cs.CL ·2026-10-07 ·Linqi Zhang, Chong Qi, Yan Cheng, Wanqing Cao, Yu Liu +2 more
benchmark
Abstract · 17 pages, 3 figures, 11 tables

Recent advances in reasoning-oriented large language models (LLMs) have motivated increasing evaluation of their ability to perform professional decision tasks. Insurance claim adjudication is one such task, requiring models to connect case evidence, insurance rules, intermediate judgments, and payout calculations across a structured decision process. We introduce InsClaimBench, an end-to-end benchmark for evaluating insurance claim adjudication across the decision chain. Grounded in real claim materials and structured insurance rules, InsClaimBench contains 3,780 cases in 375 case families across auto, property, and health insurance, comprising 86,656 atomic rule judgments. It evaluates each claim from atomic rules through adjudication modules to payout decisions and amounts, with controlled factual variants testing whether required changes are correctly propagated across levels. Evaluation of six LLMs reveals a progressive loss of reliability along the decision chain. Payout-decision accuracy ranges from 74.23--80.19%, while joint decision--amount accuracy drops to 47.54--73.15%. Strong local performance also fails to ensure case-level correctness: atomic-rule accuracy reaches 95.48%, whereas rule-vector exact match peaks at only 36.90%, and the most frequent module errors are not necessarily those most associated with final-decision failure. Under factual changes, these inconsistencies further become propagation failures: module updates are less reliable than rule updates, correct local judgments can still yield incorrect payouts, and correct payouts can conceal intermediate errors. These results show that reliable claim adjudication requires consistent composition and propagation across the decision chain.

0.36

Label-free cell counting and viability prediction with brightfield imaging and deep learning

cs.CV q-bio.CB ·2026-10-07 ·Amir Reza Vazifeh, Christian Zeigler, Sornanathan Meyyappan, Richard Jeske, Jason W. Fleischer
unusual
Abstract

Cell viability assessment is a core requirement in cell culture systems, with critical applications in biopharmaceutical manufacturing and drug development. Conventionally, it is measured by adding membrane-impermeable dyes to a sample (a process called staining), which allows compromised cell membranes to be distinguished from intact ones. However, staining has several limitations: (a) chemical agents can perturb normal cellular processes of the cells being measured, (b) it is often ambiguous to assign viability to individual cells whose membrane integrity is only partially compromised, (c) photobleaching can undermine measurement accuracy over time when using fluorescent stains, and (d) staining cannot be performed in situ or in real time. Here, we show that (1) stained cells captured under brightfield imaging contain sufficient information to distinguish live and dead cells, and (2) cells captured under unstained brightfield imaging exhibit similar image features to their stained counterparts, enabling models trained on stained cells to generalize to unstained ones. We then report the development and validation of ViabiLens, an AI-assisted software tool for label-free cell viability analysis. ViabiLens combines a cell detection model for localizing individual cells with a convolutional neural network (CNN) classifier for live/dead prediction, paired with an interactive UMAP-based viewer for visualizing and exploring individual cells across the sample. Evaluated on Chinese Hamster Ovary (CHO) cells spanning a wide range of viability conditions, ViabiLens achieves a mean absolute error of 2.33\% on unstained samples against fluorescence-based reference measurements. The demo of ViabiLens software is available at https://amirrezavazifeh.github.io/ViabiLens-Project-Page/.

0.36

Playing with Kruskal: algorithms for flat and hierarchical watershed cuts

cs.CV ·2026-10-07 ·Jean Cousty, Laurent Najman, Benjamin Perret, Deise Santana Maia
publishedunusual
Abstract

In the framework of edge-weighted graphs, watersheds have proven to be linked to well-known optimization problems, as Minimum Spanning Tree, which allowed the design of efficient algorithms for computing (hierarchical) watershed segmentations. In the present article, after reviewing the literature related to watershed segmentation, we present a detailed end-to-end pipeline of algorithms to compute (hierarchical) watershed segmentations, starting from the computation of graph-based image representations, up to the computation of connected components of the final (hierarchical) segmentation. We consider the several variations of watersheds, including their supervised and unsupervised versions, and the various ways of computing seeds, to name a few. For the first time, we bring together all these watershed notions and algorithms in a compact and understandable way. We aim at providing a reference for those interested in employing and reimplementing the watershed segmentation framework for their task at hand.

0.36

MemoCare: An Interactive Multimodal Mobile System for Automated Cognitive Screening

cs.MM cs.CVcs.HC ·2026-10-07 ·Duy-Cat Can, Mau Minh Phuc Le, Tuan-Khoa Hoang, Hai-Dang Nguyen, Trung-Hieu Do +8 more
published
Abstract · 8 pages, 1 figure, 1 table. Accepted at the demonstration track at MMM 2027

MemoCare is a React Native mobile system for automated multimodal cognitive screening in English and Vietnamese. It integrates speech, temporal and spatial orientation, touchscreen interaction, and visuoconstruction. Google Speech-to-Text transcripts are scored with deterministic, task-specific rules. GPS-grounded current and historical administrative names support location scoring, touch events verify actions, and lightweight convolutional neural networks evaluate drawings. A separate model provides visual interpretation. Software verification passed 151/151 predefined language, spatial-answer, and touch-scoring regression cases and 48/48 fixed-coordinate spatial tests across four countries. Transcript-based tests exclude speech recognition accuracy. On a locked 71-image test set, validation-selected ShuffleNetV2 x1.5 achieved 91.33% mean balanced accuracy and 78.87% exact three-criterion accuracy. Four clinician co-authors rated the workflow (pooled median 4/5 across eight criteria; item medians 3-4.5), highlighting improvements needed in speech handling, response acceptance, and result transparency. The workflow automatically produces item-level and total scores.

0.36

VideoEvolve: Co-Evolving Memory and Retrieval for Long Video Understanding

cs.CV cs.AI ·2026-10-07 ·Yongchao Xu, Bowen Ye, Jiefeng Gan, Junkai Ma, Wenzhao Li +3 more
Abstract

Long video understanding increasingly relies on external memory to organize massive visual streams into compact representations. However, most memory-based methods dynamically adapt how information is retrieved for different questions, while largely fixing what is remembered. This mismatch makes missing details costly to recover, whereas stored information is valuable only when it can be reliably retrieved. To address this issue, we propose VideoEvolve, a novel self-evolving framework that jointly evolves memory and retrieval for long video understanding. Specifically, starting from a coarse low-frame-rate overview, VideoEvolve couples a Memory Evolver for selective memory augmentation with a Retrieval Evolver for adaptive retrieval over the evolving memory. We then co-evolve the two Evolvers through alternating agentic reinforcement learning (Agentic RL), updating one while freezing the other. To steer this alternating evolution, Bottleneck-Aware Evolution Feedback (BEF) identifies whether the current bottleneck lies in memory or retrieval and directs optimization toward the more limiting side. Furthermore, VideoEvolve introduces Capability-Aware Evolution Feedback (CEF) to alleviate downstream feedback from over-specializing memory to a fixed set of training questions, shifting training toward underdeveloped yet learnable video capabilities. By integrating Agentic RL with BEF and CEF, VideoEvolve transforms downstream reasoning experience into transferable capability updates, providing a concrete path from static long-video systems toward experience-driven, self-improving multimodal intelligence. Extensive experiments on multiple long video understanding benchmarks demonstrate the effectiveness of VideoEvolve.

0.36

Counterfactual Route Optimization for Gaussian Head Avatar Modeling

cs.CV ·2026-10-07 ·Shikun Zhang, Yong Li, Yiqun Wang, Cunjian Chen
Abstract · 15 pages, 7 figures, 4 tables

Head avatar modeling requires jointly optimizing multiple objectives with different dominant effects on geometry, appearance, and cross-view consistency. However, their relative effectiveness varies across training states, while existing pipelines typically rely on fixed loss weights or handcrafted stage-wise schedules. A central challenge is therefore to identify which optimization direction is more beneficial at each training state. We propose a counterfactual route optimization framework for Gaussian head avatar modeling, which characterizes state-dependent optimization preference from the realized effects of alternative updates rather than predefined heuristic weighting. Starting from the same training state, we perform short-horizon route-restricted lookahead over geometry, appearance, and joint update routes and evaluate their outcomes under a unified utility. The resulting counterfactual evidence is factorized into a geometry--appearance preference and a residual joint advantage, separately capturing the relative preference between individual update directions and the additional benefit of coordinated optimization. We further amortize this offline evidence into a lightweight controller that directly estimates the current optimization preference and applies bounded modulation to the training objectives during full avatar optimization. Experiments on the NeRSemble dataset validate the effectiveness of the proposed design, consistently outperforming existing methods while preserving clearer local facial structures and finer details.

0.35

A Camera-Native Stereo VR180 Dataset

cs.CV cs.MM ·2026-10-07 ·Linxuan Lu
datasetcode/data
Abstract · 6 pages, 5 figures, 6 tables. Dataset: https://huggingface.co/datasets/lulinxuan/VR180

Immersive VR180 video is increasingly produced with professional stereo fisheye cameras, yet public VR180 research resources are mostly collected from online platforms such as YouTube: already stitched, projected and compressed by unknown pipelines, and without lens calibration. We present a firsthand-captured stereo VR180 dataset recorded with two Blackmagic URSA Cine Immersive cameras. It contains 1,211 samples -- 636 stereo video clips (2,220.8 s, mostly 90 fps) and 575 stereo stills -- each released as camera-native Blackmagic RAW, separate-eye native fisheye HEVC (8160x7200 per eye) and half-equirectangular HEVC (7200x7200 per eye), together with the factory lens calibration, portable fisheye/half-equirectangular conversion tools and AI-generated scene and visual-challenge annotations. Re-encoding the released fisheye and half-equirectangular renders with x265 over 24 clips, both eyes, four rate points and nine viewing directions, native-fisheye coding needed more bitrate than half-equirectangular coding at equal viewport quality for all 24 clips (median +38%), in every part of the field of view. Data: https://huggingface.co/datasets/lulinxuan/VR180 ; code: https://github.com/lulinxuan/vr180-dataset-tools

0.35

Region-Aware CLS Token Augmentation for Fine-Grained Image Retrieval

cs.CV cs.LG ·2026-10-07 ·Ian de Holanda Cavalcanti Bezerra, Vivek Trivedy, Lucas Pascotti Valem, Longin Jan Latecki
code/data
Abstract

Image retrieval methods often rely on a single global semantic descriptor extracted from an image, e.g., the [CLS] token in vision transformers. However, trying to squeeze all the semantic information of an image into a single descriptor can hurt downstream retrieval performance, especially for fine-grained retrieval tasks. In this work, we augment the semantic tokens in the newer visual transformers, the global [CLS] token and the four register tokens, with a carefully selected collection of spatial tokens, aiming to capture the spatial region representation that characterizes the contents captured in each of the semantic tokens. We leverage the DINOv2-reg model, which includes register tokens that emergently learn object and part-based representations. For each "cue" token ([CLS] and each register token), we find a "buddy" image patch token and extract an N x N patch region to produce a set of localized ROI tokens. Our approach automatically captures important regions of interest without any external bounding boxes or saliency modules, purely by matching semantic tokens with their spatial representation regions. Furthermore, we incorporate these tokens into a multi-vector retrieval framework inspired by ColBERT, enabling fine-grained matching via a per-token alignment mechanism while avoiding the large storage cost of keeping all patch embeddings. Through extensive experiments, we find that (1) register tokens encode useful fine-grained details that can complement the [CLS] token; (2) automatically pooled ROI tokens further improve fine-grained discrimination; and (3) multi-vector retrieval with a small set of tokens improves over a DINOv2-reg single-vector baseline while remaining tractable for large-scale search. The code is available at https://github.com/IdhcbIan/Augmenting_CLS_with_ROI_tokens.

0.35

On-Policy Distillation Teaches New Skills but Not New Knowledge

cs.CL ·2026-10-07 ·Yixuan Tang, Yi Yang
Abstract

On-policy distillation (OPD) strengthens language-model reasoning, yet whether students acquire new factual knowledge or compositional skill for multi-step reasoning remains unknown. We separate these capabilities using a controlled synthetic framework that measures the student's initial capabilities and independently controls the teacher's additional facts, compositional skill, or both. Across four models from three families, reverse-KL OPD reliably transfers compositional skill across unseen reasoning structures, but transfers minimal factual knowledge. Decoupling the distillation recipe reveals the source of this asymmetry: replacing reverse KL with forward KL restores factual transfer, whereas student rollouts specifically improve the execution of multi-step reasoning. Experiments on recent factual QA and competition mathematics show a similar asymmetry under reverse-KL OPD, yielding notable reasoning gains without factual memory expansion. Together, these results demonstrate that on-policy distillation does not expand a model's parametric knowledge, but instead teaches it to organize and compose the knowledge it already possesses.

0.35

Deep Learning for Detecting Building Setback Rule Violations in Urban Planning: A Case Study on Bangladesh

cs.CV cs.LG ·2026-10-07 ·Farzana Tasnim, Faria Yasmin, Md. Tauseef-Ur-Rahman, Amina Islam, Tamzid Bin Tariq +2 more
Abstract · 22 pages

Dhaka's rapid urban growth has made building setback violations common, narrowing footpaths and blocking emergency access. RAJUK, responsible for building control in the Dhaka Metropolitan Area, enforces setback rules mainly through manual field inspection, which scales poorly. We present a systematic comparison of four machine learning approaches for classifying building setback conditions from street-level photographs: negative (compliant), moderate (borderline), and positive (violation). We built a labeled dataset of 372 geo-tagged images from sixteen Dhaka neighborhoods and expanded it with 5,968 augmented copies for training. We trained and evaluated a multi-layer perceptron (MLP), a Random Forest ensemble on frozen ResNet50 embeddings, a support vector machine on frozen ResNet18 embeddings, and a fine-tuned ResNet18 CNN. On a 1,275-image test set, the fine-tuned CNN achieved 65.4% accuracy (macro-F1 0.50), followed by Random Forest at 59.9% (macro-F1 0.44) and SVM at 58.2% (macro-F1 0.45). All models classify the compliant class reliably but struggle with the borderline moderate class due to visual ambiguity. These results establish reference performance levels for automated setback-compliance screening in a South Asian regulatory setting.

0.35

Prompts versus Rules: Auditing and Controlling Speech Naturalness Behaviors in Voice User Simulators

cs.CL ·2026-10-07 ·Riqiang Wang, Elena Khasanova, Harsh Saini, Lex Konnelly, Parsa Kavehzadeh +2 more
published
Abstract · Accepted to the UserSim @ NeurIPS 2026 workshop (non-archival)

As voice agents gain more popularity commercially, the user simulators used to evaluate the deployed agents are also being developed to include more realistic, variable, and diverse speech naturalness behaviors -- disfluency, interruption and backchanneling. The quality of the user simulator directly affects the validity of agent evaluation results. However, we find that most studies so far have not examined in detail whether the intended configuration for these behaviors is realized in the simulation. In this study, we audit the realized naturalness behaviors of tau-Voice, our own LLM-based prompting approach across three models, and our rule-based injection algorithm for disfluency, interruption, and backchanneling. We find that prompting for these behaviors is unreliable and produces speech inconsistent with the instructions, placed and distributed less naturally than the instruction implies. In contrast, our rule-based, model-free algorithm produces controllable and diverse naturalness behaviors more aligned with natural speech. Our results suggest that LLMs not purpose-trained for user simulation are not sufficient on their own to represent authentic user behavior, and that linguistically informed deterministic approaches or specialized models are needed to close the gap; auditing and reporting realized naturalness behaviors, rather than configured settings, is what makes that gap visible.

0.35

HuLiGen: Human LiDAR Generation from Parametric Body Models

cs.CV ·2026-10-07 ·Salma Galaaoui, Nermin Samet, David Picard
code/data
Abstract · 12 pages, 5 figures, 7 tables

LiDAR point clouds of humans are extremely expensive to collect and annotate, thus represent a scarce resource that hinders the development of human analysis using this modality. To alleviate this scarcity, prior work relies on simulated human LiDAR, but such samples do not fully reflect the geometry and sensing characteristics of real observations. In contrast, we introduce HuLiGen, a generative model that generates human LiDAR point clouds from a parametric body model, using a point transformer trained with a flow-matching objective. We show that our generated point clouds are closer to the real capture distribution. Using HuLiGen to generate synthetic data, we propose a synthetic-only pretraining scheme for LiDAR-based HPE that achieves state-of-the-art performance, with even larger gains in low-annotation and low-data regimes, where MPJPE is reduced by up to 50%. Code, models and generated samples are available at https://github.com/valeoai/HuLiGen.

0.35

Bringing BNNs to Fast Event Processing

cs.CV ·2026-10-07 ·Paul Longour, Julien Moreau, Franck Davoine
published
Abstract · 19 pages, 3 figures, 8 tables. Accepted by NeVi Workshop at ECCV 2026

Binary Neural Networks (BNNs) enable efficient deep learning deployment on resource constrained devices with weights and activations compressed to one bit, substantially reducing model size and inference cost. Event cameras offer complementary advantages, including low latency, high dynamic range, and low power consumption, by capturing asynchronous streams of events rather than dense image frames. Despite their shared emphasis on efficiency, the combination of these technologies remains largely unexplored. This work aims at adapting and evaluating modern deep BNN architectures on event data. We also show that cross-modal pretraining from RGB data can improve the classification accuracy of BNNs on neuromorphic datasets. We introduce the Polar-wise Binary Event Volume (PBEV), a binary representation that enables event-camera data to be processed directly by BNNs and represents a step toward fully binarized event-based vision systems. Best evaluated BNN on N-Caltech101 classification benchmarks shows 90.58% accuracy with 7.5x less operations than their full-precision counterparts.

0.35

Enhancing Multi-Region Stylization with Interior-Guided Boundary Repair

cs.CV ·2026-10-07 ·Hong-Son Nguyen, Thi-Ngoc-Hanh Le
code/data
Abstract · 10 pages, 7 figures

Region-based neural style transfer enables fine-grained artistic control by allowing independent stylization of semantic image regions. However, compositing these regions often leads to boundary artifacts, degrading visual quality. We propose Interior-Guided Boundary Repair (IGBR), a lightweight and model-agnostic method that improves boundary handling in multi-region stylization. IGBR repairs boundary pixels using interior-guided propagation and applies inward, distance-based blending restricted to object-background boundaries, preventing inter-object style leakage. The method is derived from a region-wise constrained formulation with a closed-form solution and can be seamlessly integrated into existing stylization pipelines without retraining. To evaluate efficiency of our IGBR, we introduce quantitative metrics that measure boundary consistency, gradient artifacts, inter-object leakage, and interior preservation without requiring annotated stylized images. Our experiments and evaluations demonstrate that the proposed IGBR consistently produces plausible boundaries, outperforming prior blending techniques in boundary consistency, gradient stability, and interior preservation. The code is available at https://github.com/Son-SDT/IGBR.

0.34

DuRe-ST: Dual-Relation Spectro-Temporal Modeling for Speech Deepfake Detection

eess.AS cs.SD ·2026-10-07 ·Shaole Li, Siqing Qin, Youzhi Tu, Kong Aik Lee
Abstract

Previous speech deepfake detectors can adaptively capture spectro-temporal dependencies through graph attention, yet they largely overlook the co-variation between spectral and temporal representations. To address this gap, we construct a normalized affinity graph from their joint covariance and apply polynomial graph filtering to capture higher-order covariance-induced dependencies. We first develop Cov-ST to isolate the contribution of covariance-based relational modeling. Although it improves detection performance, its sensitivity to the polynomial order suggests limited robustness when covariance relations are modeled alone. We therefore propose DuRe-ST, which jointly exploits covariance-induced and graph-attention-induced relations to capture complementary second-order co-variation and adaptive spectro-temporal dependencies. Experiments show that DuRe-ST achieves an average relative EER reduction of 25.9% over XLSR-AASIST on the ASVspoof benchmarks and 28.4% across four cross-dataset benchmarks with only 4-8k additional trainable back-end parameters. It further outperforms the strongest publicly available comparison models by 2.2-13.2% in relative EER on four benchmarks, while remaining smaller than the publicly available models considered.

0.34

Less from More: Reinforcing Sparse Video Reasoning from Dense References

cs.CV ·2026-10-07 ·Wenfang Sun, Yingjun Du, Cees G. M. Snoek
Abstract

Video-language models commonly assume that more temporal observations lead to more reliable reasoning. We question this assumption and argue that the key challenge is not merely processing more video frames efficiently, but learning to reason reliably under limited temporal evidence. We propose SAVER, a dense-to-sparse post-training framework that uses dense video views as training-time references for sparse-frame inference. During reinforcement post-training, paired dense and sparse views are optimized with grounding rewards and a reliability-gated reference reward, encouraging sparse view predictions to preserve task-relevant temporal evidence. Notably, SAVER is trained only on 1,250 randomly sampled temporal grounding examples, without using any video question answering annotations. Across three temporal grounding benchmarks and six video question-answering benchmarks, SAVER consistently improves performance across frame budgets. In particular, SAVER can match or surpass dense-frame Qwen3.5 baselines while using substantially fewer frames. These results show that temporal grounding can serve as an effective evidence-localization proxy for learning sparse video reasoning that transfers to broader video understanding tasks.

0.34

AutoSynth: Learning to Generate Editable Synthesizer Programs from Audio and Text

cs.SD ·2026-10-07 ·Tristan Wu, Daniel Chin, Liwei Lin, Junan Zhang, Gus Xia
code/datahot direction
Abstract · 21 pages, including references and appendices

Audio generation models can translate natural-language descriptions into sound, but their outputs are typically waveforms. Their audio quality is constrained by audio compression, and their outputs do not readily support direct edits to notes, timbral parameters, or modulation relationships. We present AutoSynth, which represents MIDI performance events, fixed synthesizer parameters, and variable-length modulation routes as a unified sequence for a synthesizer, and learns their dependencies with an audio-conditioned autoregressive model. A single model supports both tasks. Given reference audio, the model directly predicts a synthesizer program; given text, it uses a pretrained audio generation model and converts the generated audio into a program. Training consists of two stages: supervised learning on large-scale audio-program pairs automatically constructed from a small set of native presets, followed by group-relative policy optimization with a mixed reward combining semantic similarity, pitch-related features, acoustic similarity, and sound usefulness. The pipeline requires neither paired text-target-program annotations nor a differentiable synthesizer. Experiments show that AutoSynth produces complete, editable synthesizer programs and achieves competitive results in both synthesizer inversion and text-driven generation. Audio demos and source code are available at https://auto-synth.github.io/.

0.34

BanglaRhet: Benchmarking Classical and Transformer Models for Rhetorical and Persuasion Detection in Bangla Political Speech

cs.CL ·2026-10-07 ·Rohit Kumar Sen, Anik Chowdhury
publishedbenchmarkcode/data
Abstract · 6 pages, 2 figures, 6 tables. Accepted at the 2026 2nd International Conference on Advances in Computing, Communication, Electrical, and Sma…

Political discourse often uses rhetorical and persuasive language to frame narratives, influence public opinion, and mobilize audiences. While Bangla natural language processing has made progress in sentiment analysis and opinion mining, systematic benchmarking of transformer models for fine-grained rhetorical and persuasion technique detection in Bangla political speech remains largely underexplored. This paper presents a benchmark study of transformer-based models for detecting rhetorical form and persuasive intent in Bangla political discourse. Using BanglaRhet, a manually annotated corpus of 30,289 Bangla political speech segments collected from publicly available political news sources, we formulate two supervised single-label classification tasks: rhetorical technique detection (contrast, repetition, exaggeration, metaphor, rhetorical questions) and persuasion technique detection (blame assignment, call to action, unity call, moral, emotional, and logical appeals). We evaluate four transformer-based models, BanglaBERT, BanglaBERT-Base, SahajBERT, and XLM-RoBERTa-Base, against classical TF-IDF baselines. BanglaBERT achieves the highest performance, with 65.40% macro-F1 for rhetorical technique detection and 66.46% for persuasion technique detection, outperforming the best tuned classical baseline by 19.2 and 13.8 macro-F1 points, respectively. Class-level analysis indicates that errors are mainly associated with semantic overlap among labels, figurative language, and class imbalance. The results provide initial benchmark baselines for Bangla rhetorical and persuasion-aware political discourse analysis and highlight the need for context-aware and multi-label modeling.

0.34

SpatialUQ: Post-Hoc Uncertainty Quantification from Spatial Consistency in Black-Box Vision Models

cs.CV cs.AIcs.LG ·2026-10-07 ·Md Kawsher Mahbub, Milon Biswas, Mirza Niaz Morshed, Wei Yu
code/data
Abstract

Clinical vision models are often deployed as frozen black boxes with no access to internals, retraining, or ground truth at inference time. We introduce \textbf{SpatialUQ}, a post-hoc uncertainty method using only output probabilities. It measures the Jensen-Shannon divergence between the global prediction and the mean of five fixed spatial crops in six deterministic forward passes. The premise is simple, trustworthy predictions are spatially consistent. On NIH ChestX-ray14 (DenseNet-121, $N{=}25{,}596$), our Multicrop Uncertainty Score (MUS) reaches $0.784$ failure-detection AUC versus $0.664$ for MC-Dropout ($p{<}10^{-6}$) at one-fifth the compute, with native calibration ($\text{SCE}{=}0.049$ vs.\ $0.127$ for $\ell_1$), the best-calibrated among methods above 0.78 AUC. A supervised fusion of MUS with entropy, confidence, and $\ell_1$ reaches $0.832$, outperforming a five-member ensemble ($0.813$). MUS scales with model quality, reaching $0.899$ with BiomedCLIP ($ρ= 0.846$), while this relationship remains meaningful in-distribution ($ρ= 0.523$) but breaks down under severe distribution shift (VinBigData, $ρ= 0.027$). MUS is well-suited to diffuse findings but is less dependable for small focal lesions such as nodules. Code and experimental materials are publicly available at https://huggingface.co/datasets/kawsher11/SpatialUQ.

0.34

Controllable Crowd Generation through World-Model Planning

cs.CV ·2026-10-07 ·JunGyu Lee, Jisu Shin, Seunghyun Shin, Hae-Gon Jeon
code/data
Abstract · 28 pages, 6 figures. Project page: https://jungyu0413.github.io/Ctrl-CWM

Crowd simulation plays a central role in robot navigation, autonomous driving, and urban planning. For these applications, realistic simulation requires crowds to adapt their behavior to environmental changes and user objectives. However, existing methods that rely on predefined control settings have limited flexibility in accommodating new user-specified objectives. To address this limitation, we propose Ctrl-CWM, a multi-agent Controllable Crowd World Model that integrates crowd generation and run-time control. Our key idea is to adapt the world-model principle of planning using imagined futures to crowd simulation. To this end, Ctrl-CWM consists of an encoder that learns a representation of human motion dynamics, an actor that proposes pedestrian displacements, a critic that evaluates imagined crowd trajectories, and a planner that selects actions. We first learn human motion dynamics through trajectory prediction on real-world pedestrian videos and then freeze the encoder to preserve them. Using this representation, the actor generates imagined crowd trajectories through repeated state updates, and the planner combines the critic's scores with user costs to select actions. Repeated planning advances the simulated crowd, while additional user costs introduce new control objectives without retraining. We extensively evaluate crowd generation under varied agent arrival conditions and run-time control across avoidance and attraction scenarios. Ctrl-CWM outperforms the state-of-the-art method on most crowd realism and collision metrics, and adapts crowd behaviors to user-specified objectives introduced during simulation. The project page is available at https://jungyu0413.github.io/Ctrl-CWM

0.34

The Persona Hierarchy Model: Understanding Contextual Generalization in Fine-Tuning LLMs

cs.CL cs.LG ·2026-10-07 ·Jiachen Zhao, Zhengxuan Wu, David Bau, Weiyan Shi
Abstract

Language models are routinely fine-tuned under a fixed context, such as a generic system prompt, persona or domain-specific instruction, yet the learned behavior sometimes stays confined to that context and sometimes broadly generalizes to unseen contexts. We propose the Persona Hierarchy Model to explain this: a shared default persona influences behavior across contexts. Under this model, fine-tuning that modifies the shared persona promotes broader transfer, whereas changes to local personas remain more context-specific. Across 120 fine-tuned models spanning four behaviors and 15 training contexts, generalization narrowness positively correlates with the similarity between the training context's persona and the default persona (Pearson's r = 0.72 for Qwen3-4B). Prior fine-tuning under the default context can broaden generalization in subsequent training under other contexts. Aligning contextual responses with default-persona responses produces stronger effects. Finally, we propose persona-preserving regularization (PPR) to confine undesired contextual generalization. In RL, PPR cuts reward hacking from 42-55% to at most 0.2% under every evaluated prompt while retaining accuracy gains. These results support the Persona Hierarchy Model as an explanation for contextual generalization and can motivate future controls on unintended generalization for better alignment of LLMs.

0.33

ARCS: Towards Precise Text-to-SQL via Structured Disambiguation

cs.CL ·2026-10-07 ·Yihao Hu, Yanlin Feng, Naoki Otani, Nikita Bhutani
benchmarkcode/data
Abstract

As text-to-SQL systems move beyond demonstrations toward real-world deployment, ambiguity in user questions becomes a primary source of errors. Such ambiguities are often subtle, domain- or data-specific, and can silently cause system outputs to deviate from the user's true intent. Ambiguity is traditionally addressed through conversational clarification, which is often inefficient, cognitively demanding, and poorly aligned with real-world user workflows. We propose structured disambiguation, a new paradigm in which ambiguity is resolved through explicit, constrained interactions rather than free-form dialogue. We construct ARCS (Ambiguity Resolution Corpus for SQL), the first text-to-SQL benchmark featuring naturally occurring, unconstrained ambiguities over real-world databases, with complete annotations of all valid ambiguity points, interpretations, and SQL queries. Experimental results show that text-to-SQL remains challenging in the presence of ambiguity: gpt-6-sol achieves only 51% end-to-end execution accuracy, and no open-source model exceeds 27%.

0.33

Unified Multi-plane Autoregressive Diffusion for 3D Multi-contrast MRI Synthesis

cs.CV ·2026-10-07 ·Yejee Shin, Geonhui Son, Jinglu Wang, Minwoo Jung, Yan Lu +1 more
top venue
Abstract · Accepted to ECCV 2026

Acquiring a complete set of magnetic resonance imaging (MRI) contrasts is time-intensive and uncomfortable for patients, despite the diagnostic value of multi-contrast imaging. This motivates synthesizing missing contrasts from those already acquired, which is an inherently 3D problem requiring anatomical coherence across axial, sagittal, and coronal planes. However, fully 3D generative models are often impracti- cal under computational resources that scale cubically with volume size. We propose a unified Multi-Plane Autoregressive Diffusion (MPAD), a latent diffusion framework that achieves full-volume 3D synthesis using efficient plane-wise 2D operations while preserving volumetric coherence. A 3D autoencoder first compresses MRI scans into an isotropic 3D la- tent representation. A 2D diffusion model is then trained to reconstruct masked latent slices of the target contrast, conditioned on both source- contrast slices and unmasked target-contrast slices. During inference, we introduce plane-wise autoregressive synthesis with inter-plane priors. Slices are generated autoregressively in random order within one plane orientation to maintain intra-plane continuity, then propagated as con- ditioning priors to orthogonal plane orientations to enforce inter-plane consistency. Compared to 3D latent diffusion baselines, MPAD reduces training and inference FLOPs by 7x and 3x, respectively, while also lowering inference time and peak memory consumption. Experiments on multiple datasets demonstrate that MPAD achieves superior perfor- mance, generating high-fidelity 3D volumes and supporting one-to-many translation within a single unified model.

0.33

TraceQuery: learning what to recompute from full-denoising trajectory in diffusion language models

cs.CL cs.AI ·2026-10-07 ·Yukai Jin, Zhankai Ye, Jiayue Yang, Fangyi Li, Shuoqiu Li +4 more
code/data
Abstract · 9 pages of main body, 5 pages of appendix, 4 figures, 1 table

Diffusion large language models (dLLMs) enable flexible generation through iterative masked denoising, but incur substantial inference cost because token representations are repeatedly recomputed across denoising steps. Existing cache-based methods reduce this cost by reusing intermediate results across denoising steps and recomputing only a subset of positions at each step, chosen either on a fixed schedule or by heuristic signals of change. Our trajectory experiments instead point to decoding trajectory as the more relevant quantity. Cached generation remains accurate when it stays close to the full-denoising trajectory, with errors concentrated around unmask-order deviations and reduced as longer reference prefixes are recovered. This suggests a more direct principle: limiting representation error helps, but keeping generation on the full-denoising trajectory is a more direct route to accurate decoding. We therefore formulate efficient dLLM inference as trajectory-guided querying. Rather than broadly repairing representation error, we use full-denoising trajectories to supervise which unresolved positions should receive limited fresh computation. Based on this formulation, we introduce TRACEQUERY, a lightweight sparse decoding framework that predicts a small set of near-future trajectory-relevant positions, queries only those candidates, and lets the backbone select which one to commit from their fresh predictions. TRACEQUERY delivers the strongest quality-computation trade-off among cache-based approaches, sustaining high accuracy with substantially fewer fresh token computations, and generalizes across tasks, model variants, and families. Code: https://github.com/fish1928/tracequery.git

0.33

SoccerNet-FoulRet: Retrieving Semantically Similar Soccer Foul Videos

cs.CV cs.IR ·2026-10-07 ·Jacobus Arthur, Ahmad Sait, Batool Hani, Merey Ramazanova, Jan Held +4 more
benchmarkcode/data
Abstract · ACCV 2026

Refereeing decisions in professional soccer remain inconsistent because referees cannot easily compare a contentious foul against similar past cases. We cast this as a retrieval problem and introduce SoccerNet-FoulRet, the first benchmark for semantic foul retrieval. Given a query foul, the task is to retrieve past fouls judged to be relevant precedents, regardless of camera angle, teams, or appearance. This differs from prior video-to-video retrieval, which matches clips by visual similarity or a shared event. Here, relevance is defined by refereeing interpretation. We build the benchmark from the SoccerNet-MVFoul dataset and evaluate retrieval ability of zero-shot video and vision-language embedders together with a task-specific fine-tuned baseline on 693 human-verified queries and category-relevance labels. Semantic foul retrieval remains challenging. The strongest zero-shot model achieves under 5% HitRate@10 on human-verified precedents, while category-supervised fine-tuning improves category relevance but transfers only modestly to precedent retrieval. We release SoccerNet-FoulRet to establish semantic foul retrieval as an open problem: https://github.com/SoccerNet/sn-foulret.

0.33

STLQ: Spatio-Topological Linkage Quantization for Approximate Nearest Neighbor Search

cs.CV ·2026-10-07 ·Kuizhi Liu, Junying Zhang, Xindi Ma, Xiangyu Wang, Qi Jiang +3 more
code/data
Abstract · Under review

Vector quantization is widely used for large-scale data compression and approximate nearest neighbor search (ANNS). Multi-codebook quantization (MCQ) is a practical paradigm that represents database vectors with compact codes and supports efficient retrieval through asymmetric distance computation (ADC). However, existing shallow MCQs encode vectors independently within each inverted file (IVF) cluster, without exploiting local references. Therefore, we propose Spatio-Topological Linkage Quantization (STLQ), a topological shallow MCQ method that organizes each IVF cluster as a directed-forest with compact codes and scaling coefficients. Its basic encoding phase first quantizes vectors independently, while the linked encoding phase re-encodes relative residuals with respect to selected reconstructed parents, reusing local reconstruction information to obtain finer representations. The corresponding linked-ADC preserves linear scanning complexity. STLQ further compresses the scaling coefficients and directed-forest structure to control the additional topology overhead. Extensive experiments show that STLQ achieves higher recall than the evaluated shallow MCQs at comparable bit budgets, although at a higher offline construction cost. Meanwhile, it has a competitive recall-QPS tradeoff against deep MCQs. The code is available at https://github.com/ighxiy/STLQ.

0.33

Certified by Abstention: Distribution-Free Guarantees for Chain-of-Thought Verifiers at Small Calibration Budgets

cs.CL stat.MLcs.LG ·2026-10-07 ·Arjun Balaji
publishedbenchmark
Abstract · 22 pages, 7 figures, 12 tables. Under submission at AISTATS 2027

Signals that predict whether a chain-of-thought (CoT) trace is correct are compared by AUC, but deploying one requires a threshold with a guarantee. We ask what distribution-free selective guarantees deliver for CoT verifiers at realistic calibration budgets of tens to a few hundred labelled problems, using seven open models, five verifier signals and 37,000 graded traces. The central observation is validity by abstention: an $(α,δ)$-valid procedure that issues a certificate with probability $P_{\rm fire}$ bounds the failure probability of an issued certificate only by $δ/P_{\rm fire}$, so a certificate that rarely fires can be valid and wrong every time it is used. In a simulation with known risk the standard certificate fails in at most 0.3% of calibration draws but in up to 69% of those in which it fires. A certification floor and a lattice condition for Benjamini-Hochberg conformal selection explain why certificates abstain at these budgets, and the data bear them out: the standard certificate returns nothing or a large accepted set, and an unreadable residual-stream probe buys two to three times the coverage of the readable signals, an edge a cross-fitted reconstruction cannot recover linearly from the readable features. We then give a floor-started fixed-sequence certificate, valid without monotonicity assumptions, that covers more than the Bonferroni certificate on every model-signal pair and raises coverage at the non-vacuous target $0.75π_0$ from 0.05 to 0.16, although the floor keeps absolute coverage small. Finally, a certificate cannot see what matters after deployment: under benchmark shift the error among accepted traces tracks the new task's base error, and under best-of-$n$ selection against the verifier it rises past the target while the empirical failure frequency stays below $δ$, because abstention absorbs the failures.

0.33

Do Generative Priors Align with Human Naturalness Perception?

cs.CV q-bio.NC ·2026-10-07 ·Taiki Fukiage
Abstract · 62 pages, 35 figures, including appendices

Visual generative models are trained to capture the probability distributions of natural images, yet whether their native priors reflect the regularities governing human perception of image naturalness remains an open question. Here, we probe these priors through native prediction errors across 25 open image and video generators. Because raw single-image losses are dominated by scene content and visual complexity, we evaluate directional loss differences using content-preserving, paired relational interventions that selectively disrupt facial configurations or physical illumination consistency while limiting changes in low-level image statistics. Across both domains, these loss differences reproduce human-like selective sensitivities and tolerances, capturing the classic Thatcher effect on faces and shape-dependent responses to illumination inconsistencies. Notably, these loss differences reliably track continuous gradations of human naturalness judgments across individual stimulus pairs (peaking at $r = .84$ on faces and $.64$ on physical scenes) and retain unique human-aligned signals even after controlling for feature distances from frozen vision encoders and standard image quality metrics. We also find that while overall sensitivity to these violations broadly covaries with human alignment across models, the two systematically decouple along denoising schedules, with alignment peaking earlier than sensitivity, revealing that human-like naturalness judgments dissociate from generic violation detection. Together, these findings demonstrate that learning visual distributions yields generative loss landscapes that capture distinct aspects of human naturalness perception.

0.33

Identity-Duplication Auditing in National-Scale Neuroimaging Repositories

cs.CV ·2026-10-07 ·Jiheng Li, Michael E. Kim, Trent M. Schwartz, Yuhan Cui, Gaurav Rudravaram +13 more
code/data
Abstract · Code and documentation: https://github.com/MASILab/HAPPEN

National-scale magnetic resonance imaging (MRI) repositories increasingly integrate data from different studies and institutions. However, subject identifiers assigned to study participants are typically specific to individual datasets, so the same person may appear under different identifiers when these datasets are aggregated, a problem we define as identity duplication. Such duplication can create leakage between training and test data and inflate apparent performance in downstream biomedical studies. Existing methods do not provide an end-to-end, image-based workflow to address this problem. In this work, we present HAPPEN, a human-in-the-loop pipeline for auditing identity duplication in T1-weighted brain MRI repositories. It combines SHA-256 fingerprinting for exact-duplicate detection with embeddings learned through supervised contrastive learning to retrieve near-duplicate candidates: pairs of non-identical scans that may originate from the same person. Retrieved pairs are reviewed as candidates in a locally hosted interface rather than automatically classified as duplicates. We deployed the workflow in a national-scale aggregated repository containing 95,129 scans and assessed end-to-end recovery of 54 pairs of subject records identified as duplicates using genome-wide genotype data. Transferability was assessed by locally deploying the same workflow on 22,386 scans at an independent institution without model retraining or image transfer. Deployment in the study repository identified 1,316 exact-duplicate scan groups and 1,275 near-duplicate groups of subject records supported by human review. Of these groups, 56% and 82%, respectively, crossed dataset boundaries. All 54 genetic-reference pairs were recovered. The external team independently completed the full workflow using a locally selected operating threshold and review standard.

0.32

Nobody Truly Agrees on Sentiment: Humans, Bespoke Tools, and LLMs Struggle with Social Media Texts

cs.CL ·2026-10-07 ·Himarsha R. Jayanetti, Sivakanesan Dhanushkanda, Shuai Hao, Michael L. Nelson, Michele C. Weigle
published
Abstract · 11 pages, 1 figure, 2 tables, accepted for publication at TPDL 2026

Social media is a rich source of real-time public sentiment, but widely used sentiment analysis tools are often applied without understanding their limitations. In this study, we evaluate the inter-rater reliability of three bespoke sentiment analysis tools (TextBlob, VADER, and Twitter-roBERTa-base) and three large language models (LLMs: Qwen3-32B, GPT-OSS-120B, Llama-4-Maverick-17B) against six human raters across 100 tweets. We measured agreement using two statistical measures: Cohen's kappa for pairwise comparisons and Fleiss' kappa for multiple raters. Even among the human raters, our results showed only fair agreement, highlighting the subjectivity of sentiment analysis. Higher agreement was observed under the binary sentiment classification (negative vs. non-negative and positive vs. non-positive) than under the three-class classification across both humans and automated tools. The Twitter-roBERTa-base model showed the strongest alignment with human ratings, outperforming both bespoke sentiment tools and LLMs, particularly in distinguishing negative versus non-negative sentiment. LLMs showed substantial agreement among themselves and moderate to substantial alignment with humans, performing better in positive vs. non-positive classifications. Our findings underscore that domain-specific fine-tuning remains crucial for reliable social media sentiment analysis, and human-centered evaluation remains essential for establishing gold-standard labels.

0.32

Tracing Inputs, Verifying Outputs: Validating Attribution in Music Generation

cs.SD cs.LGeess.AS ·2026-10-07 ·Taejun Kim, Wonil Kim, Jongmin Jung, Hyeongseok Wi, Sangeun Kum +8 more
Abstract · 15 pages, 4 figures. Audio examples: https://neutune.github.io/attr2027demo/

How can we verify whose music contributed to an AI-generated output? This paper demonstrates how input-based attribution can provide verifiable evidence of which audio sources were used in a generation and whether they shaped the output. To do so, we condition the generation solely on audio without any text input, then trace the inputs behind each output, and establish their musical effect. In prompt adherence tests and controlled input swaps, the stems generated by our generator, MixAudio, follow the prompt audio in timbre and the context audio in harmony. Yet these outputs may still reproduce training data not supplied as inputs. We therefore audit memorization with our musical version identification model, musicDNA, and find few reproductions outside the input records. On human-judged cases within the flagged pool, it achieves higher precision and recall than the other tested memorization detectors. The two evaluations suggest that input records and output analysis provide complementary evidence for attribution, on which rights-holder reporting and compensation can draw as the AI music economy takes shape. Audio examples are available at https://neutune.github.io/attr2027demo/

0.32

Real-Time Joint Audio-Video Generation by Parallel Adapter Composition

cs.CV ·2026-10-07 ·Jingyu Li, Xiaoxiao Xiang, Yiwen Guo
Abstract

Deploying a joint audio-video diffusion transformer for real-time, interactive generation normally requires two essential modifications: block-autoregressive attention, so frames can be emitted before the whole clip is finished, and few-step sampling, so each block is cheap. Conventionally, the streaming video literature obtains both capabilities from a chained pipeline. It first distills a bidirectional teacher into a causal student, then into a few-step one, or proceeds in reverse order. Each stage of such a chain fine-tunes the weights the previous one produced, so a later objective can undo an earlier capability. Following the idea of model merging, we show that on a packed audio-video backbone the two capabilities can be acquired in parallel. A causal adapter is trained against the frozen backbone, and an off-the-shelf few-step adapter provides the few-step capability. As the two edit different functional axes, we predict, and then verify, that their weight-update directions are near-orthogonal, without any explicit orthogonality constraint during training. Orthogonal updates should combine without interfering, so parallel composition is a direct sum. The two adapters are simply added at inference, with no joint training, yielding few-step, streaming audio-video whose image quality tracks the bidirectional teacher. Compared to the chained baselines, the composed model matches or beats them on most metrics, making parallel composition a practical approach. The resulting streaming system generates joint audio-video in real time, $\approx$26 fps at $480\times832$ without quantization, and sustains 30 s of continuous generation with stable image quality. Demo Page: https://pac-demo-2027.github.io/demo/

0.32

GRACE: Generation-aware latent compression for efficient video generation

cs.CV ·2026-10-07 ·Jiyoung Kim, Paul Hyunbin Cho, Jisu Nam, Donghoon Lee, Hyunsung Go +3 more
code/data
Abstract · Project page : https://cvlab-kaist.github.io/GRACE/, 43 pages, 24 figures

Highly compressed video autoencoders offer an effective way to accelerate video diffusion models, as the Diffusion Transformer (DiT) operates on far fewer tokens. However, such autoencoders are challenging to train, since a higher compression ratio degrades reconstruction quality and recovering it requires more channels, which is known to slow the convergence of the DiT. The compressed latent also differs from the one the DiT was trained on, so the pretrained DiT must be either retrained from scratch or adapted at considerable cost. Compressing the autoencoder the DiT was trained with appears to preserve compatibility, yet optimizing it for reconstruction alone still shifts the latent away from the distribution the DiT has learned. To address this, we propose Generation-Aware Latent Compression for Efficient Video Generation (GRACE), a two-stage framework that compresses a pretrained video autoencoder while keeping it compatible with the pretrained DiT. Specifically, we keep a frozen base latent from the pretrained encoder and learn a residual latent for the information lost under stronger compression, while aligning the compressed latent with the pretrained latent in the feature space of the frozen DiT so that the autoencoder is optimized for generation. We then adapt the DiT with lightweight fine-tuning and asymmetric denoising, where the base is denoised ahead of the residual. GRACE reduces the token count of Wan2.1-I2V-14B by 8x and its latency by 11.1x at 480x832x81, while matching the generation quality of the pretrained pipeline before compression on VBench.

0.32

Rubric Spans are Label Representations: Joint LLM Encoding for Short Answer Scoring

cs.CL ·2026-10-07 ·Zhifan Sun, Sebastian Gombert, Fabian Zehner, Leon Camus, Longwei Cong +1 more
top venuebenchmark
Abstract · EMNLP2026 Main

Automatic Short Answer Scoring (ASAS) requires models that can score student responses against question-specific criteria while remaining efficient and transferable across rubric sets. We propose RUSPAN, a rubric-conditioned ASAS framework that treats rubric descriptions as semantic label representations. RUSPAN serialises the question context, student answer, and all candidate rubric levels into a single sequence, then scores the levels listwise from the rubric-span and whole-sequence representations produced in a single LM pass. We further introduce RUSPAN-RIM, in which a Rubric-Independent Mask prevents rubric spans from attending to one another, making rubric representations depend only on the answer and question context and preventing overfitting to rubric patterns during training for zero-shot transfer. On six ASAS benchmarks spanning English, German, and Portuguese, RUSPAN improves mono-benchmark scoring over discriminative and generative baselines, while RIM with position reindexing delivers consistent and substantial gains on PT-ASAG, the held-out benchmark with the strongest combined language and rubric-structure shift.

0.32

Purifying Backdoored Large Vision-Language Models by Removing Hijacked Directions

cs.CV ·2026-10-07 ·Bojun Yang, Haochen Zhou, Zhifang Zhang, Haobo Wang, Songze Li +1 more
code/data
Abstract · 25 pages, 9 figures, 14 tables

Large vision-language models (LVLMs) are increasingly deployed in safety-critical applications, yet they remain vulnerable to backdoor attacks. Defending against such attacks remains costly, as existing methods require either extensive retraining on clean data or per-query intervention at inference time. To address this limitation, we propose OrthoPurify, a more efficient method to purify backdoored model weights via one-step orthogonal projection. Specifically, through structural analysis of backdoor weight updates, we find that the backdoor is encoded by diverting a small number of weight update directions from task adaptation to backdoor shortcut encoding, a phenomenon we term direction hijacking. However, identifying these hijacked directions requires a benign reference model, which is typically inaccessible to the defender. We show that a pseudo-benign model, obtained by fine-tuning the pretrained weights on only a small set of clean samples, provides a sufficient approximation, as the dominant update directions stabilize within the first few gradient steps. OrthoPurify uses this pseudo-benign reference to isolate the hijacked directions and removes them through a single projection on the weight update. Extensive experiments show that OrthoPurify reduces the attack success rate to near zero while preserving the original performance across diverse benchmarks, without retraining the backdoored model or introducing inference-time overhead. Our code is publicly available at https://github.com/womeimingzi/OrthoPurify.

0.31

FaceKit: a Toolkit for Interpretable Facial Phenotyping, Synthetic Image Generation and Privacy Analysis in Rare Diseases

cs.CV q-bio.QM ·2026-10-07 ·Hongzhuo Chen, Zhanliang Wang, Florent Pollet, Mian Umair Ahsan, Joshua Bie +7 more
hot direction
Abstract

Many rare genetic diseases are associated with recognizable craniofacial features. However, traditional approaches for describing facial morphology rely largely on qualitative clinical observation and free-text descriptions, which are often subjective, non-standardized, and difficult to reproduce across observers and institutions. Although the Human Phenotype Ontology (HPO) provides controlled terms for describing facial features, these terms are typically categorical rather than quantitative and may vary depending on examiner experience and interpretation. Here, we present FaceKit, a computational framework for quantitative facial phenotyping from frontal facial photographs. FaceKit extracts standardized measurements of facial landmarks and derived 120 morphological features, then reports feature-level z-scores representing deviation from population reference distributions. The reference distributions are built from the FairFace dataset spanning diverse ancestral groups. We evaluated FaceKit on a curated subset of the GestaltMatcher Database covering 50 rare-disease cohorts. In addition to quantitative facial analysis, FaceKit includes synthetic facial image generation to support rare disease model development and data augmentation. We also performed privacy evaluation to assess whether synthetic images reveal identifiable information from real patient photographs and could compromise patient privacy. Across disease case studies, FaceKit-derived quantitative measurements captured known facial features associated with rare genetic disorders and provided objective support for clinical phenotyping. Together, these results establish FaceKit as a useful tool for quantitative phenotyping, and has the potential to improve rare disease diagnosis, support genotype-phenotype studies, and enable more reproducible clinical characterization across diverse patient populations.

0.31

Masked Feature Encoding for Large-Scale Whole Slide Image Representation

cs.CV ·2026-10-07 ·Haoyu He, Basile Tessier-Cloutier, Yang Wang, Mahdi S. Hosseini
publishedcode/data
Abstract · Accepted at ACCV 2026

Whole slide image (WSI) analysis in computational pathology follows a multiple instance learning (MIL) pipeline where patch embeddings are extracted independently and aggregated for slide-level prediction, but within-slide variance from staining, scanner, and local texture can overwhelm the discriminative signal. We propose Masked Feature Encoding for Multiple Instance Learning (MFE-MIL), a feature-space masking framework that trains a lightweight MLP adapter jointly with a window-based masked reconstruction branch and a MIL classification head. The two objectives are complementary. Classification guides the adapter to suppress within-slide patch variance, while window-based masked reconstruction provides an auxiliary regularizer for the adapted features without using patch coordinates, coordinate graphs, or segmentation preprocessing. The raster patch-extraction order is used only as a weak implicit prior. At inference, the decoder is removed, leaving only the adapter and MIL head. Across CAMELYON16/17, PANDA, and TCGA-BRCA with four diverse encoders, MFE-MIL improves ACC/F1 for nearly all tested aggregator-encoder settings and AUC in most, outperforms coordinate-based spatial methods (CAMIL), and achieves higher AUC than 2DMamba on three of four datasets (UNI). On five TCGA survival cohorts it improves the average concordance index for every aggregator tested, its most consistent gain. Code is available at https://github.com/AtlasAnalyticsLab/MFE-MIL.

0.31

Iris-3B: Going Beyond the Latent with Pixel-Space Diffusion Training, Conversion and Fine-Tuning

cs.CV cs.AI ·2026-10-07 ·Hanqiu Li Cai, Chema Garabito
code/data
Abstract · 19 pages, 13 figures, 6 tables. Project lead: Hanqiu Li Cai. Code and models: https://github.com/speridlabs

Pixel-space diffusion models avoid the lossy VAE of latent models, which suggests an advantage on downstream tasks where fine-grained detail matters. We test this claim along both routes to a pixel-space backbone. We pretrain Iris-3B, a 3B-parameter pixel-space text-to-image transformer, from scratch through a $256\to512\to1024$ curriculum, after first ablating the prediction target and representation alignment at $256^2$ to decide what to scale. We also convert a pretrained latent model, FLUX.2 Klein base 4B, to pixel space. We fine-tune both families for monocular depth estimation and for image restoration/super-resolution. We find no significant improvement from using a pixel-space generative prior. Fine-tuned for depth with one matched direct-regression recipe, Iris-3B is level with the latent FLUX.2 Klein and the converted pixel FLUX.2 Klein falls behind it, and on $4\times$ DIV2K restoration neither pixel model beats a latent FLUX.2 Klein fine-tune, the converted one trailing it slightly. We document the recipes, the failure modes and the remaining confounds behind this negative result. Nevertheless, Iris-3B shows that pixel-space pretraining with the pixel-transformer (PiT) head of PixelDiT scales to 3B parameters and to text-to-image quality competitive with latent models, matching Qwen-Image on OneIG under the official evaluators at $1024^2$. We release its weights and training code in the hope that they help pave the way for further work on pixel-space generation.

0.31

MOTIP2: Spatial Priors for End-to-End Multi-Object Tracking

cs.CV ·2026-10-07 ·Benoît Roussel, Damien Bouet, Liming Chen, Pierre Perrault
published
Abstract · Accepted at BMVC 2026. 27 pages (14 pages main paper, appendix and references), 6 figures, 10 tables

End-to-end multi-object trackers have narrowed the gap with classical tracking-by-detection on association-difficult benchmarks. Yet they still make spatially implausible errors no classical tracker would, such as assigning one identity to objects on opposite sides of the frame. A model could learn to avoid them, but tracking annotations are scarce, so we encode spatial priors explicitly instead, while keeping inference fully end-to-end with no post-hoc association. We propose three spatial priors, at the data, loss, and representation stages. Spatial ID Switches bias trajectory permutations toward spatially overlapping objects, reducing the mismatch between training and inference confusions. Spatial ID Loss scales each identity's penalty by its box distance, so a distant switch costs more than a nearby one. Spatial Anchor gives each track token its frame position, an explicit spatial cue for attention. We instantiate the three priors in MOTIP2, a tracker adapted from MOTIP and built on the real-time DEIM detection transformer. Trained without extra data, its main model, MOTIP2-L, sets a new state of the art: 73.4 HOTA on DanceTrack, 76.0 on SportsMOT, and 71.1 IDF1 on PersonPath22. MOTIP2 is a family of models spanning the speed-accuracy trade-off: a lighter model, MOTIP2-S, matches the original MOTIP at over 3x the speed, and MOTIP2-X reaches 74.8 HOTA on DanceTrack.

0.31

Flow-of-Thought: A Framework for Visual Reasoning

cs.CV ·2026-10-07 ·Mariia Baidachna, Nicolas Pugeault
published
Abstract · NeurIPS WiML 2026 version: OpenReview version: https://openreview.net/forum?id=QBcqVOacYO

Mental imagery, ``seeing with the mind's eye'' is an essential aspect of human cognition. Despite rapid progress Large Language Models (LLMs) and Vision Transformers (ViTs) still underperform on tasks requiring spatial understanding. To address this, we introduce Flow-of-Thought (FoT), a framework that integrates the generation of visual sketches as intermediate reasoning steps, mimicking mental imagery in humans. We train coordinate-aware trajectory flow fields on $SO(2)$ group orbits and cumulative shortest paths, then freeze the learned dynamics; same vs. different decisions compare competing generative hypotheses using foreground-weighted reconstruction energy. On locked tests FoT reaches 100.0% accuracy on Tetris and 99.0% on colored shapes. Under frozen transfer, the orbit-trained 2D flow improves over its endpoint-only control on BLINK Multi-view (72.2% vs. 63.9% on 133 public validation pairs), supporting continuous visual traces as an effective and interpretable representation for spatial reasoning in some out-of-distribution settings.

0.31

Simulation Methods for Multiphysics Phenomena in Visual Computing

cs.GR ·2026-10-07 ·Fabian Löschner, Stefan Rhys Jeske, José Antonio Fernández-Fernández, Jan Bender
published
Abstract

Physics simulation is a cornerstone of many computer graphics applications, ranging from video games and virtual reality to visual effects and computational design. The number of techniques for physically-based modeling and animation has thus skyrocketed over the past few decades, facilitating the simulation of a wide variety of materials and physical phenomena. These course notes provide an in-depth introduction to multiphysics simulation methods for computer graphics, covering the mathematical foundations, key algorithms, and practical considerations behind the most widely used approaches. We focus on methods developed by the computer graphics community for simulating various physical phenomena and materials -- including rigid and deformable bodies, fluids, and granular materials -- as well as the interactions between them. For each method, we present the underlying mathematical framework with detailed derivations and discuss how different materials and coupling strategies fit into the formulation. A selection of software frameworks that offer out-of-the-box multiphysics modeling capabilities is also presented. Finally, we touch on emerging trends in physics-based animation, including machine learning-based methods which have become increasingly popular in recent years.

0.31

Tetris3D: 3D Scene Generation With Objects That Fit Together

cs.CV ·2026-10-07 ·Jaeyeong Kim, Jinhyuk Jang, Jongmin Lee, Kyehong Park, Seungryong Kim
datasetcode/data
Abstract · Project page: https://cvlab-kaist.github.io/Tetris3D/

We propose Tetris3D, a generative framework for single-image 3D scene reconstruction that recovers objects which are physically and geometrically coherent as a scene. Existing methods often generate objects independently or couple them implicitly, providing limited guidance for ensuring fine-grained spatial compatibility between neighboring objects that interact with one another. To address this, we explicitly condition the generation of each object on the geometry of surrounding objects and their physical relationships, guiding its shape and pose to remain geometrically and physically plausible within the scene. Moreover, we introduce ComOb, a physics simulation-based dataset of 1.2M scenes featuring physical interactions across diverse object categories, with per-object meshes and pairwise physical relation annotations. Comprehensive experiments on synthetic and realworld scenes show that Tetris3D recovers coherent object shapes and poses even when interacting regions are occluded, and achieves state-of-the-art performance in both generation quality and physical stability.

0.31

DynaConTalk: Wavelet-Constrained Diffusion for Long-Form and Controllable Holistic Co-Speech 3D Motion

cs.GR ·2026-10-07 ·Yifei Zhu, Yangyang Cai, Mingyi Shi, Miao Cheng, Lin Gu +2 more
code/data
Abstract · 14 pages, 11 figures. Code and models: https://github.com/zhuyifeiabcd1/DynaConTalk

Holistic co-speech animation is prone to averaging in both motion representation and speech conditioning. In coordinate-space diffusion, slow body posture, mid-frequency gesture strokes, and fast hand or facial details are entangled in one prediction target, often producing low-variance, over-smoothed motion. Meanwhile, dense rhythmic and acoustic cues can dominate sparse content-specific information under fixed multimodal fusion. We present DynaConTalk, a wavelet-constrained diffusion framework for long-form and controllable holistic co-speech motion generation. Diffusion operates in stationary wavelet transform (SWT) coefficient space, whose temporally aligned bands separate coarse posture evolution, gesture strokes, and fine expressive details. Our dynamic gating network preserves HuBERT and speaker identity as a base and selectively adds rhythm, mel, and transcript features through motion-state- and noise-aware residual gates. Attention pooling and learned depth routing deliver complementary conditions to each denoising stage, while a frame-resolution rhythm path preserves precise timing. A signed proposal-consensus update then reconciles these conditions with the evolving motion state. Matched-noise constraint injection uses the same sampling interface for history continuation and localized keypose repair, and extends to reference-guided control. Separate body-hand and facial denoisers, followed by inverse SWT and a pose-driven root regressor, produce holistic motion. Experiments evaluate generation quality, facial accuracy, temporal continuity, and controllable editing. Code, models, and the interactive editing interface are available at https://github.com/zhuyifeiabcd1/DynaConTalk.

0.31

MORCA: Offline-to-Online Reinforcement Learning for Adaptive Cache Reuse in Video Diffusion Acceleration

cs.CV ·2026-10-07 ·Yuxiang Xiong, Ruiyan Wang, Wenqiang Wang, Teng Hu, Songhang Shen +3 more
code/data
Abstract · 22 pages, 8 figures

Diffusion Transformers (DiTs) achieve remarkable performance in video synthesis, but their iterative denoising process suffers from high inference latency. To address this, caching has emerged as an effective acceleration strategy by capitalizing on inter-step redundancy during denoising. Existing dynamic caching methods typically estimate the error that cache reuse would introduce at each denoising step (step error) to guide cache decisions, whereas our concern is how much quality loss cache reuse would cause in the final generated video (terminal error). We show that step error does not directly correspond to terminal error and that latent information helps capture their relationship, thereby informing cache decisions. Moreover, existing threshold-based methods cannot provide precise speedup control, making it difficult to meet practical requirements for user-specified acceleration targets. To address these limitations, we introduce MORCA, a cache scheduling framework trained through offline-to-online reinforcement learning to make latent-aware reuse/recompute decisions under user-specified acceleration targets. Extensive experiments on different video generation models across multiple target acceleration ratios demonstrate that MORCA achieves better generation fidelity than state-of-the-art caching methods under comparable computational budgets. Code is available at https://github.com/x10ngyx/MORCA.

0.31

Global Average Precision for Representation Learning

cs.CV cs.LG ·2026-10-07 ·Bill Psomas, Mohammad Mahdi, Michalis Thomas, Danda Pani Paudel, Giorgos Tolias +1 more
Abstract

Standard information retrieval metrics, such as mean Average Precision (mAP), assess performance one query at a time, based on how the similarities between a query and its positives compare against those with its negatives. The same holds for common representation learning losses, such as InfoNCE and per-query AP surrogates. None of them considers whether similarities are comparable across queries, which any system with a single decision threshold relies on. Global Average Precision (gAP) does, by ranking all query-candidate pairs in one list and computing a single AP. We introduce gSAP, a differentiable surrogate of gAP. It needs only a similarity matrix and a binary matrix marking the positive pairs, the same input as existing losses, so it is a drop-in replacement for them and agnostic to the encoder, the modality, and the source of supervision. Since it considers all possible pairwise comparisons in the batch jointly, it also remains trainable at low temperatures, a regime where per-query surrogates run out of gradient. Swapping it into established recipes improves supervised metric learning, cross-modal alignment, and self-supervised pretraining, where, to our knowledge, it is the first ranking loss to replace the community standard InfoNCE in the latter two. Its similarities are more consistent across queries, which drives the gains under a universal threshold. gSAP retrieves up to four times as many positive pairs as the strongest AP surrogate at the same precision, and it degrades the least when queries with no positives in the database are added. Beyond thresholding, models trained with gSAP also learn better representations, with higher transfer, $k$NN and zero-shot classification accuracy.

0.31

Document-Level Text Simplification in Estonian Using Large Language Models

cs.CL ·2026-10-07 ·Meeri-Ly Muru, Eduard Barbu
published
Abstract · 12 pages, 2 figures, 2 tables. Published at LREC 2026

Document-level text simplification involves transformations that go beyond sentence-internal edits, addressing discourse coherence, anaphora resolution, and cross-paragraph consistency. Despite advances in sentence-level simplification for high-resource languages, document-level simplification in morphologically rich, low-resource languages such as Estonian remains largely unexplored. This study presents a comprehensive evaluation of five state-of-the-art multilingual large language models (LLMs) for document-level simplification in Estonian. Three prompting strategies are examined: single-pass generation, pipeline-based modular agents, and guideline-augmented pipelines. The evaluation framework integrates automatic metrics assessing readability, semantic preservation, and discourse coherence, alongside a structured manual annotation protocol. The findings indicate that Gemini-2.0 and LLaMA-3.3 produce outputs with near-native fluency and strong meaning preservation, whereas other models display notable grammatical and semantic limitations. This work contributes novel document-level coherence metrics, evidence-based prompting strategies, and publicly available resources for reproducibility.

0.30

Inverting Multi-Vector Visual Document Indices

cs.IR cs.CLcs.CV ·2026-10-07 ·Zhuchenyang Liu, Yao Zhang, Yu Xiao
benchmark
Abstract · 30 pages. Under review

Prevailing multi-vector visual document retrievers store each page as about a thousand patch vectors, often in vector databases run by a third party. Since no one can read a page from its vectors, this index is easily treated as less sensitive than the page. However, because the index keeps one vector per patch in raster order, and each vector is computed by a vision-language model pre-trained to read documents, we hypothesize that whoever runs or breaches the store can reproduce a page from its index alone. We frame inversion as conditional document image generation and infer from the vectors what the attack needs: the encoder, the page shape and, for shuffled vectors, their order. On the ViDoRe v3 benchmark, pages inverted from raw indices recover 47% of the words and 45% of the sensitive tokens. Used as queries against the stored indices, they rank their source page first 98.4% of the time. We test two cheap protections, token pooling and shuffling, which both cut word recall to about 8%. A model that restores the order of a shuffled index raises the share of source pages ranked first from 3.8% to 93.5%, while inverting a pooled index remains open. To test generalisation, we apply the same attack unchanged to another multi-vector retriever: its inverted pages still rank their source page first 70.2% of the time, though its word recall stays below a nearest-neighbour baseline. Multi-vector visual document retrievers are therefore vulnerable to inversion through their stored index, which should be protected like the documents it encodes.

0.30

Finding the Right Balance: Relevance and Diversity in LLM Retrieval

cs.IR cs.CL ·2026-10-07 ·Guillaume Brouillette, Faustin Kagabo, Usef Faghihi, Nadia Ghazzali
code/data
Abstract · 36 pages, 8 figures, 13 tables. Code and results: https://github.com/GuillaumeBrouillette/finding-the-right-balance

Retrieval diversification is widely available in retrieval-augmented generation (RAG) frameworks, yet prior studies disagree on whether it improves retrieval and answer quality. We show that its effectiveness varies primarily with candidate-pool redundancy, in a pattern consistent with the number of distinct evidence pieces a query requires. Using controlled near-duplicate injection and production-style overlapping chunking, we find that diversification harms relevance, evidence coverage and answer quality on clean pools, but becomes beneficial on multi-evidence tasks when redundancy causes nearest-neighbor retrieval to select repeated passages. We therefore introduce a query-adaptive rule that diversifies only when the effective number of distinct documents in the nearest-neighbor top-$k$ selection falls below the query's evidence requirement. Computed from existing embeddings, the rule captures most of the achievable gain, transfers across datasets and encoders and automatically reduces to nearest-neighbor retrieval for single-evidence queries. We also introduce RNG-Score, a geometric reranker with an exact nearest-neighbor fallback whose margin indicates duplicate structure. Overall, we conclude that diversification should be used selectively, based on observable redundancy and evidence requirements.

0.30

Masking Judder in Foveated Rendering with Phase-Shifted Noise

cs.GR ·2026-10-07 ·Ville Cantory, Danhua Zhang, Victoria Interrante, Piotr Didyk
Abstract · 11 pages, 13 figures

Low framerates during rendering arise when frame time exceeds the rendering budget, or can be introduced deliberately to save computation, for example in foveated rendering, where the periphery can be rendered at a reduced spatial and temporal resolution. However, low frame rates also cause non-smooth motion, or judder, degrading perceived image quality. Judder manifests as aliasing in the spatiotemporal frequency domain; temporal filtering, such as blurring, can suppress it, but at an additional computational cost and incur a larger loss of detail. Prior work has shown that inexpensive synthesized noise can compensate for the loss of spatial detail and velocity perception caused by foveation, but its potential for restoring motion cues lost to reduced temporal resolution has not yet been explored. We show that such noise can itself restore these cues: translating a Gabor kernel reduces to a phase shift, and modulating a kernel phase creates a perception of smooth motion without having to fully rerender the noise. We derive a method that, given the on-screen angular velocity and the temporal rendering rate, selects a spatial frequency noise band such that phase shifts perceptually mask judder without introducing motion aliasing of their own. User experiments show that our technique substantially expands the range of angular velocities over which motion at reduced temporal resolution remains smooth, without objectionable degradation of image quality.

0.30

GHARP: Real-time Gaussian Head Animation from Large-scale Reconstruction Prior

cs.CV cs.LG ·2026-10-07 ·Ali Benlalah, Sepehr Johari, Patricia Vitoria, Armin Kappeler, Artem Sevastopolsky +12 more
publishedbenchmark
Abstract · Accepted to ACCV 2026. 35 pages (14 main + references + 15 pages supplementary), 14 figures, 17 tables

We present GHARP (Real-time Gaussian Head Animation from Large-scale Reconstruction Prior), a method that animates 3D human heads in real time from a few input images of a subject and a driving expression signal. We decouple the problem into an identity stage that builds a representation of the subject's geometry and appearance offline, and an animation stage that predicts expression-dependent residuals on top of it at runtime. This separation offers a favorable trade-off with respect to fidelity, quality and runtime: the identity stage can be expensive while the animation stage runs a lightweight network, optimized for mobile devices. Our method performs animation in a semantically structured latent space of a pretrained reconstruction model, where expression changes remain spatially contained, making residual prediction efficient. This reconstruction prior provides a consistent spatial layout, allowing fusion of multiple input views into a compact, fixed-size canonical Gaussian representation. While this two-stage design improves the runtime-quality trade-off, it still inherits a problem common to all expression-driven avatar methods: expression codes describe only the face and thus omit body pose and clothing position, making these regions underspecified in the input. The animation network faces an ill-posed mapping and resorts to averaging over conflicting body appearances, producing blur and temporal flicker. We address this with a body alignment network that learns to align the person's body in the target image with the input reference images, removing the ambiguity from the training signal. Our method achieves state-of-the-art quality on the Ava-256 benchmark while running up to 13x faster on an A100 GPU with 8x fewer Gaussians.

0.30

UltraWorld: Learning Interactive Ultrasound World Models from Untracked Clinical Videos with Acoustic Sampling Map

cs.CV cs.RO ·2026-10-07 ·Keke Yang, Erqi Wang, Sainan Guan, Hongliang Ren
code/data
Abstract

World models can enable autonomous ultrasound scanning by predicting the outcomes of probe motions from local observations. Learning this action--observation relationship typically relies on synchronized video--pose pairs, which are costly to collect at scale and largely unavailable in routine clinical recordings. Reliable action following further requires modeling ultrasound's cross-sectional sampling geometry. We present UltraWorld, a self-distillation recipe that transfers priors from clinical ultrasound videos into interactive world models without real action annotations. Starting from clinical videos, we adapt a video foundation model into an ultrasound generator conditioned on reference images and anatomical masks. Anatomical masks sampled along programmable trajectories through 3D anatomy provide spatial guidance for synthesizing action--video pairs. We then use these synthetic pairs to self-distill the generator into a world model that predicts future observations from local observations and actions, without requiring anatomical masks or other 3D assets at inference time. To further improve action following, we introduce the Acoustic Sampling Map (AsMap), which represents probe poses and imaging settings as pixel-wise 3D sampling positions, beam directions, and depths. Experiments demonstrate improved prediction fidelity and action following. Across nine simulated closed-loop local planning episodes, UltraWorld reduces the mean final distance to the goal and orientation error by 29\% and 38\%, respectively, compared with visual servoing. Project Page: https://ultraworld-project.github.io/.

0.30

PPCAR-Net: Projection-Refined Parametric 3D Coronary Artery Reconstruction from Sparse X-ray Angiographic Views

cs.CV ·2026-10-07 ·Yu Ren, Hwee Kuan Lee, Tat-Jen Cham, Jonathan Yap, Khung Keong Yeo
code/data
Abstract · 22 pages, 5 figures, 10 tables, including references and appendix. Code and pretrained models: https://github.com/G2304138H/PPCAR-Net . Proj…

Sparse-view 3D coronary reconstruction commonly relies on cross-view correspondence and triangulation, which are vulnerable to vessel overlap and foreshortening, or on volumetric prediction followed by vascular-graph extraction, which does not directly provide centrelines and radii. We introduce PPCAR-Net, a projection-refined parametric coronary artery reconstruction network that directly predicts a branch-structured centreline-and-radius representation without explicit point matching, triangulation, or an intermediate volume. Given a variable number of segmented views, a coarse predictor combines frozen VGGT features with learned branch queries to estimate branch presence, B-spline centreline trajectories, and dense radius profiles. Projection-guided geometry and radius refiners then sample local evidence from the input views and apply residual corrections learned with 3D supervision. We evaluate representation fidelity and sparse-view reconstruction quantitatively and qualitatively. On simulated angiographic masks generated from CT-derived coronary anatomy, PPCAR-Net produces better connected artery reconstructions and achieves strong centreline accuracy, particularly for RCA, while maintaining competitive volumetric overlap. Coarse-to-fine inference takes 121 ms, enabling real-time reconstruction.

0.29

CASM: Context-Aware Semi-Markov Post-Processor for Beat Tracking

eess.AS cs.SD ·2026-10-07 ·Zhanhong He, Hanyu Meng, Yaolong Ju
Abstract · Submitted to ICASSP 2027 conference. Code and demo: https://zhanh-he.github.io/casm-beat-tracking

In music beat tracking, model-predicted activations must be decoded into a discrete, musically coherent event sequence. Direct peak picking closely follows local evidence but can retain spurious peaks or miss weak beats. Dynamic Bayesian networks (DBNs), a widely used structured post-processor, improve sequence consistency under predefined global tempo, meter, and transition constraints, but their behavior can depend strongly on these settings. We introduce CASM, a context-aware semi-Markov decoder that instead conditions its temporal constraint on local activation evidence. CASM also accounts for ambiguity among competing periodic interpretations, including half- and double-tempo alternatives. Deterministic safeguards prevent implausible outputs and preserve beat-downbeat consistency. Applied to fixed activations from three neural beat trackers (BeatThis, MSCNN, and TCN), CASM improves temporal continuity while preserving event-level F1 across the GTZAN and SMC datasets, without backbone retraining or dataset-specific retuning. Further analysis shows that CASM is less sensitive than the DBN baseline to the composition of the calibration data.

0.29

InstanceBench: Diagnosing Referential Reasoning and Target Identity in Referring Expression Segmentation

cs.CV ·2026-10-07 ·Yuchen Li, Shaoyang Zhou, Yiran Wang, Ruiyi Deng, Haoyu Wang +3 more
benchmark
Abstract

Referring Expression Segmentation (RES) links natural-language descriptions to pixel-level object masks. Yet standard evaluation provides limited insight into instance-level referential reasoning: it does not systematically distinguish referential logics, test target preservation across valid grounding paths, or separate target-selection from mask-generation errors. We introduce InstanceBench, an instance-centered diagnostic benchmark comprising 6,194 images, 9,264 target instances, and 25,077 human-verified expressions. Each target-centric expression set (TCES) fixes the image and target mask while pairing a minimal expression with a same-target variant that uses another valid cue or grounding path. A compact referential-logic taxonomy spans direct target evidence, same-class selection, relational and compositional grounding, and exclusion, while logic-critical construction suppresses simpler shortcuts. Identity-aware metrics measure target retention and set-level success while separating selection from mask-generation errors. Across 22 native-mask RES checkpoints from 18 model families, the strongest checkpoint reaches 67.1% mIoU but only 59.6% All@0.7. Controlled interventions confirm language sensitivity, while failure decomposition identifies target selection rather than mask decoding as the main bottleneck. On a controlled training subset, matched supervision improves identity-aware performance, showing that the diagnosed capability responds to targeted supervision. Collectively, InstanceBench supports a measure-diagnose-improve cycle: measuring target consistency across grounding paths, localizing failure sources, and evaluating targeted interventions.

0.29

SAPD: Step-Aligned Privileged Distillation

cs.CL ·2026-10-07 ·Tianle Wang, Jiayu Liu, Ruizhi Zhao, Ning Miao
code/data
Abstract · preprint

On-policy post-training can improve large language models by learning from their own trajectories, but requires costly rollout generation. We ask whether fixed demonstrations can support competitive off-policy learning through better supervision. Our premise is that their usefulness depends not only on the training trajectories, but also on whether supervision provides informative preferences among continuations and connects this guidance to the reasoning decision being learned. We introduce Step-Aligned Privileged Distillation (SAPD), a rollout-free self-distillation method that turns demonstrations into step-aligned distributional supervision. Its key insight is to use the known progression of a reference solution to associate each reasoning transition with targeted privileged guidance, rather than treating the solution as undifferentiated context. On mathematical reasoning benchmarks, SAPD outperforms supervised fine-tuning and label smoothing on average while remaining competitive with on-policy reinforcement learning and self-distillation. Analyses support both the value of context-dependent distributional guidance and the benefit of aligning privileged information with the current step. SAPD also largely preserves out-of-domain coding performance and achieves approximately 2x training-loop speedups over the on-policy baselines. These findings suggest that carefully constructed supervision can make fully off-policy post-training a competitive and computationally efficient alternative. Our code is available at https://github.com/Miaow-Lab/SAPD.

0.29

SpikingVLA: Asynchronous Spiking Vision-Language-Action Models

cs.CL ·2026-10-07 ·Jingya Wang, Dehao Zhang, Shuai Wang, Malu Zhang, Yang Yang +1 more
Abstract

ANN-to-SNN conversion offers a practical route toward energy-efficient spiking Vision-Language-Action (VLA) models by bypassing the substantial cost of training large-scale SNNs from scratch. However, existing methods often require many timesteps to maintain competitive performance, resulting in substantial inference latency for real-time VLA deployment. To address this challenge, we introduce SpikingVLA, an ANN-to-SNN conversion framework that enables accurate and low-latency spiking VLA inference. Specifically, we propose a Dendritic Integrate-and-Fire (DIF) neuron that alleviates channel-wise activation outliers through dendritic mixing and adaptive somatic firing, enabling accurate ANN-to-SNN conversion with fewer timesteps. Building on DIF neurons, we further introduce an asynchronous execution mechanism that overlaps temporal computation across VLA components, reducing synchronization overhead and latency. Extensive experiments demonstrate that SpikingVLA achieves competitive navigation performance with substantially improved inference efficiency. Compared with existing spiking VLA methods, SpikingVLA improves SR and SPL by 11.9\% and 12.6\%, respectively, while reducing first-action latency by 11.2$\times$. These results establish SpikingVLA as a practical framework for deploying pretrained VLA models with high-performance and low-latency spiking inference.

0.29

VERDICT: Visual Extraction-Result Decoder with Integrated Confidence and Traceable Evidence

cs.CV cs.LG ·2026-10-07 ·Harikrishnan P M, Goutham Vignesh, Siva Prasad, Zishan Ahmad, Mahesh Basveshwar Deshmukhe +3 more
published
Abstract · Accepted at NeurIPS 2026, Workshop on VLM4RWD

Enterprises run schema-guided structured field extraction on documents with frontier multimodal models, and accuracy keeps improving as new models and systems arrive. Two things have not kept pace: the evidence for each extracted value, meaning the location of the document it was read from, and how far to trust the value and that evidence. Neither is a component of most extraction systems today, so we build them as a framework that attaches to an existing extractor and leaves it untouched. We present VERDICT, a plug-in framework that uses a small Vision--Language Model (VLM) to ground/localize the evidence once per document over a fixed external extractor's output, then calibrates value and evidence confidence. Fine-tuning the VLM for evidence localization substantially improves visual evidence quality and, more importantly, expands VERDICT's selective-prediction operating range. At a strict confidence threshold $0.80$, VERDICT with the fine-tuned VLM gives usable precision--coverage trade-offs for both value and evidence confidence across the in-distribution and out of distribution datasets, whereas the frozen setting often collapses to little or no coverage. Running an open-weight VLM for localization rather than a closed API makes its internal states available to the calibrator, and they are worth more than output log-probabilities: in-distribution AUROC rises from $0.730$ to $0.914$ for value confidence and from $0.681$ to $0.906$ for evidence confidence. This advantage does not survive a distribution shift. We show that the signal is displaced rather than destroyed, and that labeling roughly $120$ fields from the target dataset takes evidence confidence from $0.641$ to $0.905$ on one dataset and from $0.675$ to $0.829$ on the other.

0.29

Toward AI-Driven Phytoplankton Event Detection from Image Classification: Temporal Aggregation Under Domain Shift

cs.CV ·2026-10-07 ·Ali Rohan, Hassan Shayan, Han-Gue Jo, Colin Moffat
unusual
Abstract · 22 pages, 7 figures

Imaging FlowCytobot (IFCB) systems generate large volumes of phytoplankton imagery that exceed the capacity of manual expert analysis. Although machine-learning approaches achieve high image-level classification performance, it remains unclear whether their predictions can provide ecologically meaningful temporal information for monitoring. This study investigates whether AI-derived phytoplankton classifications can reproduce temporal occurrence patterns and elevated-occurrence signals under realistic sampling and independent-domain transfer. A monitoring-oriented framework was developed using MedPlanktonSet and evaluated on an independent IFCB dataset from the Swedish Meteorological and Hydrological Institute (SMHI). Eight biologically comparable taxa were selected as targets, while all remaining observations were retained as background to preserve the full image-stream denominator. Session-aware partitioning prevented temporal leakage, and alternative preprocessing strategies and convolutional neural network architectures were evaluated. Cross-domain experiments showed that preprocessing strongly affected generalisation. EfficientNet-B0 provided the strongest target-versus-background discrimination and was selected for temporal analysis. Despite residual image-level errors, AI-derived target-occurrence rates closely reproduced expert-derived signals after aggregation (mean absolute error = 0.0432, Pearson $r$ = 0.9696, Spearman $ρ$ = 0.9568). Agreement remained strong across multi-session windows, and elevated, strong, and persistent occurrence proxies were recovered in held-out testing. These results show that AI-based phytoplankton classification can support monitoring-relevant temporal indicators beyond image-level identification and support future operational early-warning development.

0.29

Scalable Patch-Level Self-Supervised Learning

cs.CV ·2026-10-07 ·Maximilian Seitzer, Gabriele Trivigno, Antonín Vobecký, Seungeun Yi, Maxime Oquab +3 more
Abstract

Self-supervised learning (SSL) at scale produces powerful visual representations. However, most scalable SSL methods rely on ad hoc combinations of multiple objectives and stabilization mechanisms. Taking a step back, we ask if we can design a high-performing, yet principled SSL algorithm. Starting from the multi-view assumption, stipulating that task-relevant content is captured by the information common to different views, we construct an information-theoretic objective decomposing into interpretable terms. This derivation yields JEM, a student-teacher method that learns by aligning corresponding patch representations across views, explicitly regularized by information and structure preservation losses. JEM trains stably from 300M to 7B parameters, and, to our knowledge, is the first latent-space patch-level method demonstrated at 7B scale. Across all scales, JEM reaches strong performance on both global and dense probing tasks, on segmentation benchmarks consistently surpassing the DINOv2 algorithm, an influential foundation for today's strongest visual SSL methods. Notably, at 7B parameters, it exceeds the performance of DINOv3 on panoptic segmentation, despite being trained on $12\times$ less data without refinement stages. These results demonstrate that we can indeed design an SSL algorithm that learns strong representations, is principled and stable.

0.29

Enabling Preference-driven Unlearning in Few-step Distilled Text-to-Image Diffusion Models

cs.CV cs.LG ·2026-10-07 ·Gaurav Patel, Jun Fang, Greg Ver Steeg, Qiang Qiu, Sravan Sripada
top venue
Abstract · Accepted at NeurIPS 2026

Text-to-image diffusion models are increasingly distilled into few-step variants and being deployed to enable fast inference. However, their ability to generate harmful or undesired content poses significant safety risks. Data-driven unlearning methods suppress targeted generations by fine-tuning model weights using specialized unlearning objectives. Crucially, these objectives implicitly rely on multi-step denoising dynamics, an assumption that breaks down for few-step distilled (FSD) models, resulting in ineffective forgetting. Furthermore, performing unlearning on the non-distilled base model and subsequently re-distilling it to obtain an unlearned FSD model incurs substantial computational and time overhead, making it impractical in many settings. Hence, we address this limitation with a preference-driven unlearning framework that revisits Direct Preference Optimization (DPO) for diffusion models. We show that standard DPO and its unlearning derivatives, formulated around noise-prediction error, transfer poorly to FSD models due to their altered generation dynamics. To overcome this, we introduce a modified preference optimization formulation explicitly aligned with the few-step generation properties, enabling direct concept removal in FSD models while preserving few-step efficiency and maintaining strong retention of desirable (non-targeted) capabilities. We evaluate our framework primarily on identity and NSFW (nudity) removal tasks and also extend our method to object-level unlearning. Extensive experiments demonstrate consistent and effective forgetting, and strong retention performance, establishing our method as a practical and principled solution for unlearning in FSD models.

0.28

Position Forcing: Self-Conditioning 3D Generation

cs.CV ·2026-10-07 ·Ziheng Ouyang, Zeqiang Lai, Jiarui Chen, Jiangshan Wang, Yuhao Wan +5 more
hot direction
Abstract

Recent single-stage 3D generative models commonly adopt VecSet representations, encoding 3D shapes as unordered sets of latent tokens. However, compared with two-stage methods that provide explicit positional guidance, these models must implicitly infer token positions throughout denoising, limiting their generation quality. We observe that, despite the absence of explicit positional conditioning, VecSet tokens retain recoverable spatial correspondences. Building on this observation, we propose Position Forcing, a position-based self-conditioning framework. During denoising, Position Forcing recovers token positions from the current clean latent estimate, quantizes them at progressively finer resolutions according to the denoising stage, and feeds the resulting positional encodings back into the diffusion Transformer. This progressively refined positional feedback provides spatial guidance at a granularity appropriate to each denoising stage, guiding shape generation along a coarse-to-fine trajectory and substantially improving generation quality without a separate position generation stage. Experiments demonstrate that Position Forcing achieves strong performance among single-stage 3D generative methods and outperforms several competitive multi-stage approaches.

0.28

SkillCycle: Co-Evolving Agent Policies and Skill Banks

cs.CV ·2026-10-07 ·Ling Li, Qiuyu Shen, Zheng Jiang, Qinwei Ma, Yuxuan Liu +1 more
hot direction
Abstract

Internalizing external skills changes a language agent's capabilities and, with them, the value of its remaining guidance: rules can become redundant, misleading, or insufficient for newly encountered decisions. This creates a coupled problem of learning from skills and adapting the skills that supervise further learning. We introduce SkillCycle, a framework for co-evolving agent policies and skill banks through a feedback loop between skill internalization and rule revision. Our central contribution is to give distillation feedback a second role: token-level contextual differences help locate rules for inspection, while interaction outcomes guide edits to their content and applicability. SkillCycle alternates between two phases: policy learning with a fixed skill bank and router, and rule revision with a frozen policy. Candidate edits undergo rule-level and whole-bank environment comparisons before they guide the next learning cycle. On WebShop, SkillCycle with a 3B model achieves a success rate of 74.74% and a score of 88.37 without inference-time skill inputs, representing relative improvements of 0.73% and 3.96% over the state-of-the-art (SOTA) model, respectively. In Cycle 3 ablations on ALFWorld and WebShop, SkillCycle's no-skill success rates improve by 10.18% and 18.11% relative to a static skill bank, and by 2.41% and 2.50% relative to a single bank update, respectively. These results show that continually revising skill guidance as the agent's capabilities change helps transform external skills into policy capabilities that require no skill inputs at inference. We will release code, configurations, skill banks, and evaluation protocols.

0.28

Beyond Group Splits: Specimen-Level Cross-Validation and Visual Attribution for Remaining-Shelf-Life Regression in Climacteric Fruit

cs.CV ·2026-10-07 ·Rovhona Mudau, Jean Frederic Isingizwe Nturambirwe, Clement Nthambazale Nyirenda
Abstract · 7 pages, 1 figure, 4 tables

Estimating remaining shelf life (RSL) from images could provide affordable decision support for perishable produce, but evaluation protocols can substantially affect reported performance when repeated images are available from the same biological specimen. We use the Hass Avocado Ripening dataset, comprising 8,834 image-RSL pairs from 426 fruits across three storage regimes, to evaluate a frozen ImageNet-pretrained visual backbone with a lightweight regression head. Our contributions are threefold: we quantify the effect of observation-level versus specimen-disjoint evaluation, compare lightweight and heavier visual backbones under specimen-disjoint cross-validation, and examine their spatial attributions using Grad-CAM. Across ten observation-level random splits, the model achieves a mean RMSE of 2.37 days with a standard deviation of 0.03 days, whereas specimen-disjoint 5-fold cross-validation yields a mean RMSE of 3.12 days with a standard deviation of 0.11 days. The corresponding mean coefficient of determination is 0.553. A matched per-specimen comparison confirms higher error under specimen-disjoint evaluation, with a probability value below 0.001 across 426 specimens, showing that observation-level partitioning gives a substantially more optimistic estimate for this dataset and model configuration. Under specimen-disjoint evaluation, MobileNetV3-Small (0.93 million parameters) achieves accuracy comparable to ResNet-18 while providing substantially higher throughput, and Grad-CAM reveals differences in spatial attribution between the lightweight backbones. These results support specimen-disjoint evaluation and attribution analysis when assessing lightweight vision models for longitudinal shelf-life prediction.

0.28

GraphRectify: Graph-Based Transfer of Adversarial Example Detectors Across Neural Networks

cs.CV ·2026-10-07 ·Arash Vashagh, Roozbeh Razavi-Far
code/data
Abstract

Adversarial example detectors are often tied to the classifier backbone they were trained on, limiting reuse when the protected model is replaced or upgraded. Directly transferring such detectors across backbones is challenging because different networks generally produce incompatible internal representations. We propose GraphRectify, a graph-based framework for transferring adversarial image detectors across classifier backbones. GraphRectify learns a structured representation of intermediate classifier features and adapts representations from a new backbone to the detector learned on the original model, enabling detector reuse. We evaluate GraphRectify across multiple datasets, backbone architectures, and adversarial attacks, including detector-aware adaptive attacks that jointly target the classifier and detector. Across the complete evaluation matrix, GraphRectify achieves higher aggregate ROC-AUC than training a detector from scratch on the new backbone and the evaluated transfer ablations. The gains are particularly strong for transfers between different backbone families and when sufficient data are available. In contrast, training from scratch remains competitive in the most data-limited settings. These results show that adversarial detection knowledge can transfer effectively across heterogeneous classifier architectures rather than being relearned whenever the protected backbone changes.

0.28

LLM4Impact: Integrating Heterogeneous Information for Scientific Impact Prediction

cs.CL ·2026-10-07 ·Yong Cao, Markus Flicke, Haoyu He, Katrin Renz, Andreas Geiger
benchmark
Abstract · 27 pages, 12 figures, 11 tables

Predicting the future impact of a newly published paper is challenging because it must be inferred from heterogeneous evidence available at publication time. Existing approaches often rely on a single source of information or combine multiple sources without accounting for their different predictive roles. In this paper, we present LLM4Impact, an evidence-aware method for scientific impact prediction that learns to represent, integrate, and calibrate heterogeneous information. LLM4Impact combines semantic, graph, LLM, and temporal representations, and injects graph information into a frozen LLM through continuous prefix tokens. A context aware gating mechanism adaptively weights different evidence, while a separate calibration module accounts for domain and temporal variation in citation scales. We further construct a large-scale benchmark dataset with 2 million papers, leakage-safe point-in-time heterogeneous ego graphs, temporal splits, and both year-level and month-level citation targets. Experiments show that LLM4Impact consistently outperforms strong semantic, graph, and LLM based baselines, with a 10.13% reduction in year RMSE on the in distribution test set and a 6.87% reduction under out-of-domain distribution. Our results reveal that the value of such evidence is context dependent: different papers benefit from different sources, while domain and publication time affect how evidence translates into citations. This finding motivates adaptive evidence selection and context-conditioned calibration rather than simply richer representations. We will release our code, benchmark, and an interactive web demonstration upon publication.

0.28

CIRSeg: Coarse-to-Fine Intensity-Robust Liver Segmentation with Source-Free Continual Test-Time Adaptation

cs.CV ·2026-10-07 ·Ruoshi Xu, Mingqi Gao, Shengda Luo, Jingkun Chen
publishedcode/data
Abstract · Accepted at the CARE 2026 Workshop at MICCAI 2026

Reliable liver segmentation in contrast-enhanced MRI is essential for quantitative hepatic assessment, treatment planning, and longitudinal disease monitoring. However, limited annotated data and scanner- or vendor-dependent intensity variations can cause overfitting and poor generalization to unseen acquisition domains. Moreover, simultaneously achieving robust global localization and precise boundary delineation remains challenging, while predictions may contain isolated false-positive regions outside the main liver component. To address these challenges, we propose CIRSeg, a coarse-to-fine, intensity-robust liver segmentation framework based on nnU-Netv2. CIRSeg combines 3D CutMix with stochastic intensity transfer using either Nyul augmentation or histogram matching to improve robustness to heterogeneous MRI intensities. Its cascaded architecture decouples low-resolution anatomical localization from full-resolution boundary refinement. At inference, source-free test-time adaptation based on confidence-filtered predictions and probability-prior regularization further improves robustness to out-of-distribution inputs. As a final deterministic post-processing step, largest connected component filtering removes isolated false-positive regions. On the CARE 2026 test set, CIRSeg achieves Dice scores of 97.13\% and 97.93\% on the in-domain and unseen-domain subsets, with corresponding HD95 values of 20.18 mm and 11.30 mm, respectively. These results demonstrate consistently accurate segmentation across both in-domain and unseen acquisition settings. The code is available at https://github.com/jingkunchen/MICCAI_CARE_2026

0.28

Large Language Model-Assisted Preparation of Transportation Management Plans: A Case Study with WisDOT WisTMP System

cs.CL ·2026-10-07 ·Zihao Sheng, Pei Li, Zilin Huang, Yen-Jung Chen, Yuhao Luo +4 more
dataset
Abstract

Work zones are critical yet hazardous components of transportation infrastructure, requiring carefully designed Transportation Management Plans (TMPs) to ensure safety and mobility. However, TMP preparation remains labor-intensive and heavily dependent on practitioner expertise. This paper proposes a Large Language Model (LLM)-assisted framework to automate TMP content generation, leveraging the WisDOT WisTMP system as the application context. The framework fine-tunes multiple open-source LLMs across different model scales and deploys them locally to ensure data security. To support model training, we construct a domain-specific dataset from historical WisTMP documents by converting PDF files into structured question-answer pairs in JSON format. Experimental results show that fine-tuning significantly improves performance across standard text generation metrics. Further section-wise and strategy-level analyses reveal that, while LLMs achieve strong overall performance, they tend to over-generate strategies and struggle to produce project-specific justifications and accurate cost estimates. In addition, scaling from 7B/8B to 14B yields limited gains. These findings demonstrate the potential of LLMs to improve TMP preparation efficiency while highlighting remaining challenges in LLM-assisted TMP development. The source code and demo videos will be publicly available at https://zihaosheng.github.io/TMP-LLM/.

0.27

When to Unpair: Regulating Pairing Dependence in Medical Visual In-Context Learning

cs.CV ·2026-10-07 ·Cheng Wan, Chenjun Li, Qingyu Zhao
Abstract · 24 pages, 12 figures

Visual in-context learning (ICL), well suited to label-scarce medical imaging, uses support image-label pairs to demonstrate input-output mappings, while the labels collectively indicate the requested task. We diagnose dependence on individual pairings with a test-time derangement that reassigns every support label to another support image while preserving the query, support images, and label multiset. The resulting pairing gap, defined as shuffled-minus-matched performance, shows that all four released models depend on the pairing, to widely varying degrees. Further analysis of a paired-trained model reveals support-associated spurious regions and lesion-size biases even with real, unaltered supports, alongside sensitivity to mis-registered support labels. To regulate this dependence, we introduce a late unpairing curriculum (LUC), which starts with matched training and then applies random unpairing, replacing each support label with that of another support in the same episode. LUC nearly closes the pairing gap on two backbones while maintaining or improving matched-support performance across all evaluated task types, with gains extending to held-out tasks and cross-dataset episodes. It also mitigates these failure modes. On BraTS whole-tumor segmentation, matched-support DSC rises from 0.733 to 0.857 while the gap shrinks from -0.184 to -0.008. In a released model, brief fine-tuning with random unpairing reduces the gap. A reversed curriculum that places the same number of unpairing epochs at the start of training leaves a large gap. This shows that pairing dependence is shaped by the order of training and not only by the amount of unpaired training.

0.27

Stochastic Graph Compression for Constant-Memory Differentiable Light Tracing

cs.GR ·2026-10-07 ·Linas Beresna, Eugene Fiume
top venue
Abstract · 4 pages. SIGGRAPH Asia 2026 Technical Communications

We describe Reservoir Light Replay Backpropagation (ResLRB), a method for reverse-mode differentiable light tracing whose memory is constant in path length and in the number of sensor connections per light path. Naive automatic differentiation of a light tracer records a computation graph that grows with the number of valid sensor connections along each path, since flux can be splatted from every scattering vertex; adjoint path replay removes the analogous dependence on depth for viewpoint path tracing, but its non-branching structure does not extend to the splatting case. We compress the adjoint graph during the primal pass by stochastically retaining a single representative sensor connection per light path via a streaming weighted reservoir. The adjoint pass reconstructs the path deterministically by replaying its pseudorandom number sequence and backpropagates only through the retained connection, yielding an unbiased gradient estimator whose peak memory is that of differentiating a single scattering event.

0.27

How Do LLMs Change Predictions Under Negation?

cs.CL cs.AI ·2026-10-07 ·Jongwook Yoon, Jongwon Lim, Sungjib Lim, Woojin Cho, Yohan Jo
benchmarkcode/data
Abstract · Under Review

Negation is an essential feature of human language, yet large language models (LLMs) remain unreliable in processing it. We evaluate recent open-source and closed-source LLMs on our negation benchmark and find that, in 37-71% of cases, they repeat the same answer under negation (e.g., "Madrid" for "What is not the capital of Spain?"). To understand and address this brittleness, we mechanistically examine how models operate under negation. Our main finding is that specialized attention heads and MLP neurons jointly implement negation by (1) suppressing retrieval of the original answer (e.g., "Madrid") while (2) promoting a favored candidate within the answer category (e.g., "Paris"). This contrasts with accounts of human negation processing, in which information about the original answer helps to determine what should be excluded. Furthermore, we find that this difference from human processing is a key source of negation failures: the model's mechanism relies on suppressing the original answer rather than using it to determine what to exclude, so the model can repeat the original answer when suppression is too weak or when a bias toward particular answers prevents it from selecting an alternative. To address this weakness in the model's negation mechanism, we propose a training objective that requires larger shifts in answer preference for more confident original predictions, and show that it reduces negation failures with less degradation of general capabilities than standard fine-tuning baselines. Together, our results demonstrate how mechanistic analysis can reveal why a linguistic capability fails and guide training that targets the underlying limitation.

0.27

Gaussian Material Fields for Volumetric Multi-Energy CT Decomposition

cs.CV ·2026-10-07 ·Jian Lin, Jiancheng Fang, Hongming Shan, Shaoyu Wang, Yang Chen +1 more
Abstract · 13 pages, 12 figures

Volumetric material decomposition in multi-energy computed tomography requires a representation that organizes multiple three-dimensional material fields in a common spatial domain while retaining differences in composition and local structure. We observe that spatial primitives can be shared across materials without tying their coefficients, but their local capacity must respond to material-specific reconstruction needs. We introduce Gaussian material fields, which represent multiple material distributions with shared anisotropic 3D Gaussian primitives and independent nonnegative material coefficients. The shared geometry defines a continuous spatial basis, while the coefficients determine each primitive's contribution to the individual material fields. To reconstruct this representation from multi-energy projections, a differentiable spectral forward model combines Gaussian material path integrals with a calibrated basis matrix, enabling joint optimization of spatial geometry and material composition. Material-aware adaptive density control retains material-specific refinement evidence before aggregation and adjusts local representation capacity to accommodate both spatially extensive components and sparse details. Experiments use synthesized multi-energy projections generated from pseudo-reference material maps constructed by conventional methods from publicly available CT data. Across 15 cases, our approach improves average PSNR by 4.03 dB and SSIM by 4.96% over the strongest baseline, while reducing NRMSE by 33.45%. Material-wise comparisons and component ablations support improved recovery of localized structures, while runtime and memory measurements show favorable computational scaling. These results establish Gaussian material fields as an explicit, adaptive representation for volumetric multi-material reconstruction.

0.27

TRACK: Telemetry-Based Racing Analysis and Coaching Kit in Sim Racing Games

cs.GR cs.AIcs.LG ·2026-10-07 ·Efe Çangırılı, Murat Kurt
unusual
Abstract · 29 pages, 9 figures, 8 tables

This paper presents TRACK (Telemetry-Based Racing Analysis and Coaching Kit), which is a framework for analyzing driving performance in sim racing and profiling how individual drivers behave behind the wheel. We report this framework together with its limitations: we calibrate each clustering result against a null, and when one does not separate from chance, we say so. Instead of restricting ourselves to scoring drivers or sorting them into preset labels, we represent each recording session as a compact geometry in a four-dimensional behavioral space (speed, braking, strategy, and consistency), and we group these fingerprints by their similarity using unsupervised clustering. Over time, we have developed and refined this framework on the open Assetto Corsa Gym (ACGym) dataset. Our study suggests that corner types differ along a behavioral dimension that was not used to define them. It also suggests that when the car changes, only speed and consistency carry over in the restricted population, while repeatability could not be shown there for any of the braking or strategy measures. Cluster separation becomes less distinct as the range of available telemetry widens. Until that repeatability is shown, grouping on the braking and strategy dimensions cannot treat the car as interchangeable, which divides an already small sample into smaller cells. It is also not clear whether a driver's grouping carries over from one corner type to the next. We also normalize each metric against a reinforcement-learning reference agent. The reference does not depend on the sample, so the scale does not shift when the sample does. We intend these results as an analytical foundation for a personalized improvement suggestion system. The sample is small. The cross-car result changes when the sample is defined more broadly. These outcomes are preliminary.

0.27

I would rather quit NLP than read another paper like this: The rise of antithesis in NLP papers

cs.CL ·2026-10-07 ·Olga Zamaraeva, Adrián Gude, Roi Santos-Ríos, Carlos Gómez-Rodríguez
unusual
Abstract · 30 pages, 2 figures, 48 tables

For better or worse, LLMs are by now used routinely for scientific writing.\footnote{This paper is no exception; we did use AI to assist with writing some of the sections (see Acknowledgments).} Many have noticed that recent models fill papers with unnecessary antithesis, stating over and over what the work does not do, in ways that do not contribute to its precision or quality of expression and annoy reviewers \emph{rather than impressing them}. We study the construction \emph{rather than} in ACL papers from 2019, ACL-style arXiv papers from 2026, and papers written by GPT models from the same titles and abstracts. Its rate in 2026 is seven times the 2019 rate, and higher still in the GPT papers. Two annotators, blind to the source, find almost no 2019 use \emph{annoying} and about one in ten 2026 uses; they seldom agree on which, yet about half of 2026 papers contain a use that annoys each of them. \emph{Annoying} uses present the rejected alternative less favorably than legitimate uses. Raters of preference data and open reward models favor the construction, and an instruction to be honest promotes it. We conjecture that it is a side effect of post-training on pairwise preferences, which credit a disavowal in a single response and cannot register its cost across a text.

0.27

HeiCo-FOCUS: A Clinically Grounded Dataset for Long-Context Video Understanding

cs.CV ·2026-10-07 ·Leon Mayer, Lucas Luttner, Patrick Godau, Kai Fritzsche, Annika Reinke +40 more
benchmarkdatasetcode/data
Abstract · 28 pages, 9 figures, 5 tables. Code: https://github.com/IMSY-DKFZ/orena-focus

Recent advances in Vision-Language Models (VLMs) have led to rapid progress in video understanding across a wide range of benchmark tasks. However, existing evaluations largely focus on short-term reasoning, failing to assess a critical capability: maintaining cumulative temporal consistency over extended time horizons. To close this evaluation gap, we introduce HeiCo-FOCUS, a clinically grounded dataset for evaluating long-context video understanding through the task of Foreign Object Contextual Understanding in Surgery. Built on a dataset of Heidelberg Colorectal surgeries, this task requires models to continuously track multiple objects as they are inserted, manipulated, occluded, and removed over procedures lasting up to hours. HeiCo-FOCUS comprises 30,000 visual question answering (VQA) pairs covering five core capabilities: object recognition, temporal grounding, aggregation, event and procedural understanding, and complex reasoning. The dataset was constructed through a rigorous multi-stage annotation pipeline involving large-scale crowd annotation and 39 surgical domain experts to ensure high quality and clinical relevance. To systematically probe model behavior, we introduce a multi-track evaluation framework that progressively increases temporal and contextual demands from single frames to full procedures. Experiments with ten frontier VLMs show that HeiCo-FOCUS tasks are far from solved: only around half of the models clearly outperform a text-only baseline. Across the video tracks, models perform best on event and procedural understanding (mean Accuracy: 56.5% across all models), while temporal grounding remains particularly challenging for all evaluated models (mean Accuracy: 19.7%). We therefore expect HeiCo-FOCUS to serve as a catalyst for the development of models capable of reliable, temporally consistent reasoning over hours-long videos.

0.27

MIRA: A Musical Intent Refinement Agent for Aligning Text-to-Music Generation with User Intent

cs.SD cs.AI ·2026-10-07 ·Zekai Liu, Zhilin Wang, Xuzheng He, Yu Cheng, Yang Yang
benchmarkcode/data
Abstract

Text-to-music systems produce increasingly convincing audio, yet evaluation reveals little about whether the result matches user intent. A global text-audio relevance score can overlook the implicit intent in underspecified prompts and mask failures in specific requirements, such as instrumentation, structure, rhythm, or mood progression. To bridge this gap, we formulate text-to-music intent alignment as satisfying a per-request rubric of independently verifiable items covering both a request's explicit requirements and its implied musical intent. Scoring items individually makes evaluation diagnostic by intent source and musical dimension, rather than a single opaque score. We instantiate this as MuRA-Bench, a benchmark of real-world platform requests curated by music experts. We further propose MIRA (Musical Intent Refinement Agent), a test-time agent that first grounds a request's intent into rubrics, then searches over prompt revisions for a black-box generator under a bounded budget, iteratively generating music, verifying it against the rubrics, and using this feedback to guide a trajectory-aware tree search. Experiments across open-source and commercial backends show that MIRA improves intent alignment, enabling an open-source generator to achieve performance comparable to representative commercial systems (e.g. Suno and Mureka). Project page: https://mirareview.github.io/.

0.26

Hardware-aware Calibrated Clustered Attention for Efficient Visual Geometric Transformers

cs.CV cs.LG ·2026-10-07 ·Weitian Wang, Shubham Rai, Cecilia De La Parra, Akash Kumar
published
Abstract · Accepted to IJCNN26

The Visual Geometry Grounded Transformer (VGGT) marks a significant leap forward in 3D scene reconstruction, as it is the first model that directly infers all key 3D attributes (camera poses, depths, and dense geometry) jointly in one pass. However, this joint inference mechanism requires global attention layers with extremely long sequences that causes a significant latency bottleneck. In this paper, we propose blockwise clustered attention (BC attention) to accelerate the global attention layers in VGGT. By limiting the clustering within HW-friendly neighborhood blocks, BC attention reduces the computation overhead of query clustering as well as the costly data movement between on- and off-chip memory. This enables BC attention to scale to long sequences and deliver practical latency improvements on GPUs. Moreover, we introduce a hashing hyperplane calibration method and a threshold-based error compensation method to reduce clustering errors efficiently, which is a bottleneck in the current clustered attention mechanism. Overall, our experiments on GPU demonstrate that calibrated BC attention accelerates the global attention layers by 2.10-2.63$\times$ and the whole backbone by 1.77-2.35$\times$ with negligible loss (1%) for large scenes. With a small performance loss (< 5%), calibrated BC attention further achieves a 2.26-2.87$\times$ latency improvement on the global attention layers and a 1.90-2.55$\times$ improvement on the backbone.

0.26

$Δ$Representation: Geometry Supervised Representation Learning of Phenotypes via Counterfactual Reasoning for Medical VLMs

cs.CV ·2026-10-07 ·Hao Wang, Qiwei Zeng, Jinghao Lin, Shuchang Ye, Yuezhe Yang +4 more
code/data
Abstract

Medical vision-language models (VLMs) have shown increasing potential for radiological image interpretation. Medical VLMs encode radiological images into visual representations that capture both anatomical and phenotypic information for diagnosis. Existing approaches improve pathological phenotype representations through semantic-guided representation alignment. However, pathological phenotypes arise as lesion-specific visual changes superimposed on underlying normal anatomy. Such semantic alignment approaches fail to model the phenotype-specific increment relative to the corresponding normal anatomical representation. To address this gap, we propose \textbf{$Δ$Representation}, a visual phenotype representation learning framework based on counterfactual reasoning for medical VLMs. It comprises \textbf{BaseAnatomy}, a geometry-supervised representation learning module, and \textbf{$Δ$Phenotype}, a counterfactual incremental representation learning module. BaseAnatomy provides fine-grained geometric supervision through spatial relationships across and within anatomical structures. $Δ$Phenotype computes the representation increment between lesion representations and their corresponding normal anatomical representations, and supervises increments associated with the same phenotype to cluster in the representation space. Experiments on \textit{ReXGroundingCT} and \textit{LIDC-IDRI} demonstrate that $Δ$Representation effectively structures pathological phenotype representations and improves lesion grounding and phenotype characterization accuracy in medical VLMs. Code is available at https://anonymous.4open.science/r/deltarep-CF6D.

0.26

HySPE: Positional Encoding via Symplectic Dual Shears

cs.CL ·2026-10-07 ·Zhongping Ji
Abstract · 11 pages

We introduce Hyperbolic Symplectic Positional Encoding (HySPE), grounding positional attention in non-compact symplectic transformations. While canonical Rotary Position Embedding (RoPE) parameterizes the compact, elliptic branch of $\Sp(2,\R)$ via rotations, HySPE operationalizes its hyperbolic branch via a damped symmetric composition of dual shears, yielding a conformally symplectic contraction with two spectral decay rates per channel pair. To eliminate the exponential representation drift inherent to naive absolute factorizations, we diagonalize the operator in its invariant eigenbasis and introduce blockwise coordinate rebasing with adaptive centered execution. This guarantees length-independent numerical bounds while matching cached RoPE forward latency (7.21\,ms on an RTX 4090). On TinyShakespeare, HySPE-UltraLong maintains an invariant perplexity of 4.810 up to $16\times$ zero-shot extrapolation ($L=4096$), whereas RoPE degrades to 131.198. Scaled to a 51M-parameter subword Transformer on WikiText-103 ($L_{\text{train}}=512$), HySPE closely matches RoPE in-domain while robustly extrapolating to length 8192, reducing tail perplexity by 83.9\% over RoPE. While these controlled experiments establish HySPE's extrapolation robustness and numerical stability, evaluating its scaling behavior on large-scale foundation models remains an important direction for future investigation.

0.26

From Pixel to Coding: Evaluating the Figure Reproduction Capabilities of MLLMs

cs.CV cs.AI ·2026-10-07 ·Zijian Chen, Zhengyu Chen, Bohan Liang, Lirong Deng, Yushuo Zheng +5 more
code/data
Abstract · 46 pages, 18 figures

Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in both visual understanding and code generation. However, existing benchmarks typically evaluate these two modalities in isolation, lacking a dedicated assessment of their unification, i.e., how a model can perceive complex visual structures and synthesize them into precise, executable code. Moreover, current visual code generation benchmarks often rely on simplified layouts within single programming environments, falling short of evaluating true unified multimodal reasoning. To bridge this gap, we propose FigCodeBench, a comprehensive framework for rigorously evaluating MLLMs on figure reproduction, integrating multimodal comprehension and generation. We first design a systematic dataset construction pipeline, resulting in a total of 6,194 instances that cover 7 functional categories and 4 types of programming languages. We further categorize figure reproduction into three tiers with visual and code complexity modeling, specifically targeting complex structural reasoning, varying aspect ratios, and dense geometric constraints. We introduce a multi-dimensional evaluation protocol, encompassing visual fidelity and syntactic isomorphism, that aligns highly with the Mean Machine Opinion Score (MMOS) and human preferences. Based on our framework, we conducted extensive experiments on 24 widely used proprietary and open-source MLLMs (e.g., Gemini 3.1 Pro, GPT-5.4, and Kimi-K2.5), where we observed a universal, non-linear performance cliff across different programming languages and difficulty scenarios for all models, and gained several insights, such as the significant metric decline in rigid declarative languages.

0.26

Arctic Questions, Missing Answers: A Dataset and Benchmark for LLM Abstention in Arctic Science

cs.CL cs.AI ·2026-10-07 ·Benjamin Wilcox, Dawei Gao, Pradeeban Kathiravelu, Douglas Causey, Kewei Sha +1 more
benchmarkdatasetcode/data
Abstract

Large language models (LLMs) should abstain from scientific multiple-choice questions when no option is valid, but frequent abstention alone does not demonstrate sensitivity to answer availability. We introduce ArcticQA, a dataset of 194 questions derived from primary Arctic research, with automated checks of answer support and distractor contradiction against source evidence. We further develop ArcticAbstain, a paired benchmark comparing answer-present and answer-absent conditions, with the correct answer replaced by a distractor in the latter and an explicit abstention option in both. We evaluate eight models from the Gemini, Claude, and ChatGPT families at high reasoning effort, with three trials per condition, yielding 9,312 recorded responses. Answer-present abstention rates range from 0.0% to 63.0%, whereas replacing the correct answer increases abstention by 5.05 percentage points on average. These findings highlight substantial baseline differences and the need to evaluate abstention frequency and responsiveness jointly. The dataset and benchmark are available at https://github.com/BenWilcox8/arctic-qa.

0.26

An Invariant Tangent-Angle Descriptor and a Band U-Net for 2D Fragment Adjacency Prediction

cs.CV cs.LG ·2026-10-07 ·Guillaume Brouillette, Alain Goupil, Pierre-Olivier Parisé, Fadel Touré
benchmark
Abstract · 14 pages, 4 figures, 6 tables

This paper addresses the prediction of adjacency between pairs of 2D fragments based on their contours. We improved the two-stage architecture proposed in Beaulac's thesis, in which a rotation-equivariant Siamese convolutional neural network scores pairs of local image windows along the two contours of two fragments. The scores are gathered in an adjacency matrix in which a ResNet detects the partial anti-diagonal band that reveals the adjacency of two fragments. In the current work, we keep the pipeline and replace the local score by a comparison of tangent-angle profiles of contour windows, making it, by construction, invariant to fragment rotation and agnostic to the selected contour-starting point. These adaptations may be either a training-free likelihood ratio or a small one-dimensional convolutional model trained on corresponding points. We also replaced the final classifier by a band U-Net that segments the band and classifies the pair, so that the shared arc is obtained along with the decision. In the synthetic data set of the original thesis, the tangent descriptor performs as well as or better than the image-window approach in all tested configurations. The proposed pipeline reaches an accuracy of 98%, vs 93% to 95% for the original approach once its evaluation is corrected. We tested our pipeline, with models trained only on synthetic data, on the PairingNet benchmark, and obtained an AUC of 0.93. Furthermore, under the PairingNet pair-searching protocol conditions, our learned descriptor obtains a Recall@10 of 0.82 on the real set against 0.56 from the best model of the original paper.

0.26

Self-correction Optimization for Interleaved Multimodal Generation

cs.CV ·2026-10-07 ·Xin You, Zhiwei Ning, Zukai Chen, Minghui Zhang, Xuanke Shi +5 more
Abstract · 20 pages, 10 figures

Multimodal large language models (MLLMs) have made significant progress in visual understanding and generation. However, generating interleaved image--text content remains challenging, as it requires tightly integrated multimodal understanding and generation capabilities. Although existing MLLMs provide promising solutions, most rely on additional training with augmented data, which is computationally expensive and remains limited in preserving visual subjects, temporal consistency, and physical plausibility. In this work, we propose self-correction optimization (SCO), an effective training-free method for consistent interleaved generation. SCO treats the classifier-free guidance update as a reference and performs minimal self-correction under two complementary constraints, including new-event and state-preserving constraints. Specifically, the new-event constraint promotes temporal consistency across image--text sequences, while the state-preserving constraint maintains the coherence of visual subjects throughout subsequent generation steps. Experiments on challenging interleaved multimodal generation benchmarks demonstrate significant improvements in temporal coherence and visual-subject preservation. Furthermore, SCO can be extended to video generation and improves the modeling of physically grounded processes, including robot manipulation and long-horizon handcrafting.

0.26

BEANS-Next and ROOTS: Broadening Audio-Language Capabilities for Bioacoustics

cs.SD cs.LG ·2026-10-07 ·Christos Plachouras, David Robinson, Marius Miron, Gagan Narula, Paul Laisné +15 more
benchmarkcode/data
Abstract · 36 pages, 7 figures. Project page: https://earthspecies.github.io/beans-next/

Bioacoustics and ethology encompass a wide range of audio understanding tasks, many of which stand to benefit from recent advances in large audio-language models. However, progress in the field has so far been assessed on a narrow set of tasks, primarily centered on label-centric biological category recognition, such as species and call-type classification. In this work, we introduce BEANS-Next, a benchmark grounded in a taxonomy of bioacoustics tasks spanning acoustic perception, biological category recognition, scene understanding, and in-context learning. Using BEANS-Next, we show that existing models exhibit limited performance beyond the task families emphasized by existing evaluations, constraining their usefulness for broader bioacoustic applications. To support progress on this broader task space, we also introduce ROOTS, a large-scale training resource built from expanded curated real-world data and previously underused behavioral and acoustic metadata, supplemented by audio-derived information and scalable synthetic generation where labeling is insufficient. We demonstrate that training on this dataset yields substantial progress across all task groups of BEANS-Next, moving audio-language models closer to their potential as general-purpose assistants for bioacoustics and ethology. To accelerate progress in the field, we open-source our benchmark, dataset, and data pipelines.

0.26

TileSkipper: Region-Adaptive Tile Pruning for 3D Gaussian Splatting

cs.CV ·2026-10-07 ·Jingxing Li, Yongjae Lee, Deliang Fan, Abhay Kumar Yadav, Cheng Peng +1 more
Abstract

Tiled 3D Gaussian Splatting rasterizers often use one scene-wide contribution cutoff for tile enumeration, although content differs in its sensitivity to support truncation. TileSkipper selects a static per-Gaussian cutoff policy for a frozen checkpoint. Calibration renders measure candidate pair savings and an isolated-removal distortion proxy that accounts for front transmittance and background color. The method allocates cutoffs across 64 Gaussian groups and accepts policies only after complete renders on disjoint selection views. The exported policy uses one byte per Gaussian, with no parameter updates, additional kernel, or per-frame policy inference. Across 13 scenes from Mip-NeRF 360, Tanks & Temples, and Deep Blending, a fixed-policy AccuTile sweep gives dataset-macro speedups of $1.088\times$ at standard resolution and $1.238\times$ at 3840 pixels wide, with $-0.007/-0.023$ dB mean PSNR change. Six integrations with existing opacity-aware bounds yield $1.009\times$--$1.121\times$ compiler-only speedups. For four ports from $3σ$ rasterizers, we separately attribute the prior exact-bound transition and our incremental gain. Matched-quality ablations show modest gains over scene-global calibration and parity with per-Gaussian control; the standalone comparison with AdaGScale is regime-dependent.

0.26

Diffusion-Generated Image Watermarking: A Two-Axis Taxonomy and Three Protocol-Bounded Case Studies

cs.CV ·2026-10-07 ·Sung Ju Lee, Nam Ik Cho
published
Abstract · 17 pages, 2 figures. Accepted to the non-archival track of the ECCV 2026 LifeGenIP Workshop. English translation with partial reorganization…

Watermarking diffusion-generated images requires balancing provenance signals with image quality, robustness, and computational cost. This work organizes methods along two axes: insertion mechanism and primary signal-bearing representation, and formalizes a representative $z_T$-Fourier pipeline for verification and identification. We then use the taxonomy to structure three protocol-bounded case studies. The first examines associations among frequency integrity, detection, quality, and cropping behavior. The second revisits persistence under seed-linked and seed-independent editing and formulates a scoped Semantic Imprinting Hypothesis without claiming a localized carrier or causal mechanism. The third studies single-shot VAE-latent phase modulation, including its efficiency, regeneration robustness, and robustness--quality operating points. Finally, we separate four content-level attack families from model/pipeline adaptation, propose corresponding evaluation protocols and testable conjectures for parameter-tuning threats, and identify additional temporal extensions for video. These analyses do not establish a universal ranking; instead, they provide a framework for matched, protocol-aware comparisons of watermarking systems for diffusion-generated images.

0.25

Direction-preserving Spatial Audio Signal Enhancement in The Spherical Harmonic Domain

eess.AS ·2026-10-07 ·Huawei Zhang, Jihui Aimee Zhang, Huiyuan June Sun, Prasanga Samarasinghe
published
Abstract · Accepted by the 55th International Congress & Exposition on Noise Control Engineering (INTER-NOISE 2026)

Spatial audio signal enhancement (SASE) aims to separate a desired soundfield from interference and noise while preserving its spatial cues. However, existing algorithms pay limited attention to preserving the spatial cues of the residual interference, which reduces their effectiveness in practical applications. In this paper, we propose a direction-preserving SASE algorithm based on multi-output beamforming. The proposed algorithm extracts the desired soundfield while preserving the spatial cues of both the desired soundfield and the residual interference soundfield. The simulation study shows that the proposed algorithm achieves performance comparable to the existing SASE algorithm in terms of estimation error, while it better preserves the spatial cues of the residual interference within a region of interest, in a noisy and reverberant environment.

0.25

STORK: Spatio-Temporal Observation of uterine contRactions via neural networKs

cs.CV ·2026-10-07 ·Melissa Schween, Tristan Gottwald, Jordina Aviles Verdera, Lisa Story, Mary Rutherford +1 more
published
Abstract · Accepted at the PIPPI Workshop at MICCAI 2026 and will appear in the workshop proceedings (Springer)

Uterine contractions in fetal MRI are typically identified manually and discarded, limiting insights into contraction dynamics. We formalize Uterine Contractile Activity Detection (UCAD) as a weakly-supervised learning problem and introduce STORK, a multi-instance learning model trained on dynamic MRI series using only coarse, series-level labels. STORK factorizes 3D spatio-temporal convolutions into parallel branches across temporal hyperplanes to capture coherent tissue motion without the cost of full 4D convolutions. Per-frame embeddings, combining intensity and Demons-estimated displacement fields, are aggregated by a linear mean-pooling head. This ensures that frame-level contraction scores can be recovered post-hoc without frame-level training supervision. Evaluated on around 700 multi-vendor dynamic fetal MRI series, STORK achieves a series-level AUROC of 95.0% and AUPRC of 94.6%, substantially outperforming 3D ResNet and ConvNeXt baselines. Grad-CAM analysis suggests that the model draws on predictive features extending beyond the placenta into the uterine tissue, offering an automated tool for richer phenotyping of uterine behavior.

0.25

TiTok: Audio-Visual LLM for Multi-Segment Temporal Grounding

cs.CV ·2026-10-07 ·Eunji Shin, Dahyun Choi, Seungyeon Jo, Yejin Hong, Jiyoung Lee
code/data
Abstract · ACCV'2026

Audio-visual multi-segment grounding (AV-MSG) in untrimmed videos, reasoning over audio-visual evidence and predicting multiple segments for a query, is a fundamental problem but remains challenging. Visual-only models overlook complementary acoustic cues, while audio-visual models often fail to calibrate the number of events - a phenomenon we refer to as count miscalibration. We present TiTok, an audio-visual large language model (AV-LLM) that localizes an arbitrary number of temporal event segments for each query. For precise boundary prediction, we introduce the Time Token Interleaving (TTI) method, which explicitly injects special time tokens into the audio-visual stream to align input-side temporal perception with output-side temporal prediction. We further propose decoupled, multi-segment-oriented rewards for reinforcement learning, consisting of global, local, count, precision, and format rewards, optimized with Group reward-Decoupled Normalization Policy Optimization (GDPO). To assess the performance on AV-MSG, we establish a new UnAV-100-based evaluation protocol, and propose the CountF1 metric for quantifying count miscalibration that overlap metrics fail to capture. TiTok reaches 65.7 mIoU and 0.58 CountF1, achieving state-of-the-art performance. Our code is available at this link.

0.25

Benchmarking Adult Addressee Classification Across Child- and Adult-Directed Speech Datasets

eess.AS ·2026-10-07 ·Sofoklis Kakouros, Daniil Kocharov, Okko Räsänen
published
Abstract · Accepted at SLT 2026

In this work, we present a comprehensive analysis of classification performance for distinguishing child-directed speech (CDS) from adult-directed speech (ADS) using speech data from corpora containing natural in-lab and in-the-wild CDS and ADS. We establish classification benchmarks for these datasets using self-supervised learning (SSL) representations, along with a range of time-pooled representations that go beyond first- and second-order statistics by incorporating cross-channel covariances in high-dimensional embeddings. In addition, we probe these representations to examine how different pooling methods capture prosodic information using linear probes. Overall, SSL-based representations prove particularly effective, achieving the best performance, while different pooling methods offer complementary advantages for the task.

0.25

BagDINO: Multi-View Baggage Re-Identification with DINOv3

cs.CV ·2026-10-07 ·Vita Santa Barletta, Danilo Caivano, Rebecca Margiotta, Massimiliano Morga, Davide Pio Posa
benchmarkunusual
Abstract · 7 pages, 4 figures, 3 tables, IEEE International Conference on Evolving and Adaptive Intelligent Systems 2026 (IEEE EAIS 2026)

Mishandled checked baggage remains a recurrent issue in airport operations, and current recovery workflows still largely rely on tag-based tracking, which does not directly support visual identification when tag evidence is missing or unavailable. This paper investigates baggage re-identification as an instance-level retrieval problem in a multi-camera setting, leveraging DINOv3 foundation-model representations to match a query image against a gallery of registered baggage images. A Torchreid-style BNNeck re-identification head is placed on top of a DINOv3 backbone, and parameter-efficient adaptation is performed via LoRA. Experiments are conducted on the MVB benchmark using a progressive study that compares a fully frozen backbone against LoRA and fine-tuning strategies. Results indicate that parameter-efficient adaptation of foundation-model features provides an effective and stable approach for multi-view baggage re-identification under limited training data.

0.25

Geometry-Supervised Visual Representation Learning for Multi-Phenotype Lesion Interpretation in Medical VLMs

cs.CV ·2026-10-07 ·Hao Wang, Qiwei Zeng, Shuchang Ye, Jinghao Lin, Yuezhe Yang +3 more
code/data
Abstract

Medical vision-language models (VLMs) have shown increasing potential for clinical image interpretation. However, these models still struggle to interpret multi-phenotype lesions whose diagnosis requires the joint assessment of multiple pathological phenotypes. Existing vision-language alignment methods produce visual representations that fail to preserve anatomical hierarchies and relationships among phenotypic subclasses. This stems from their reliance on semantic supervision, which lacks geometric constraints to preserve these relationships in the visual embedding space. Moreover, the sparsity of lesion-related anatomical and phenotypic representations makes it difficult for medical VLMs to capture important diagnostic evidence. To address these limitations, we propose \textbf{PureVision}, a geometry-supervised visual representation learning framework for multi-phenotype lesion interpretation in medical VLMs. It combines a geometry-supervised representation learning module, \textbf{PureEyes}, and an anatomy-guided evidence aggregation module, \textbf{PureNeurons}. PureEyes provides geometric supervision through ideal spatial distributions that encode anatomical hierarchies and phenotypic subclass relationships. PureNeurons projects visual representations into the learned latent space, using their positions to selectively aggregate lesion-specific anatomical and phenotypic evidence. Experiments on \textit{LIDC-IDRI}, \textit{CBIS-DDSM}, and \textit{3DReasonKnee} demonstrate that PureVision improves lesion grounding and phenotype characterization in visual question answering and radiology report generation. Code is available at: https://anonymous.4open.science/r/purevision-06C2.

0.25

Rubix: Global Correspondence-Free Point Set Alignment through Assignment Geometry

cs.CV cs.CGcs.LGcs.RO+1 ·2026-10-07 ·Subhransu S. Bhattacharjee, Dylan Campbell, Rahul Shome
theory
Abstract · 67 pages, 20 figures. Includes full proofs and experimental appendices

Procrustes-Wasserstein alignment jointly estimates a matching and rotation without supplied correspondences, but alternating minimization can stop at suboptimal solutions. Rubix solves the equally weighted planar problem globally under squared Euclidean loss. Each matching $σ$ of two centered $n$-point sets defines a complex correlation $z_σ=\sum_i\bar x_i y_{σ(i)}$. Their convex hull is the permutation polygon: supporting vertices give optimal matchings at fixed rotations, and the farthest vertex gives the global alignment. We prove the sharp bound of $n(n-1)$ vertices for $n\ge2$, answering Rote's rotation-assignment open problem. In exact arithmetic, assignment queries recover the polygon in $\mathcal O(n^5)$ operations. Assignment-based bounds extend the approach to three-dimensional rotations and partial matching at a supplied translation through branch-and-bound. On timed MPEG-7 shape pairs, Rubix attains every numerical reference value in 12 ms on average, 50 times faster than a rotation grid at the same accuracy. Its distances improve gravity-aligned matching of real 3D scans, shape retrieval and noisy crystal classification over alternating minimization.

0.25

CrossEdit: Cross-Modal Training Enables Rich Audio-Visual Editing

cs.SD eess.ASeess.IV ·2026-10-07 ·William Chen, Prem Seetharaman, Ke Chen, Oriol Nieto, Kevin Duarte +8 more
benchmark
Abstract · 27 pages (10 main + 1 limitations); 8 figures

Multimedia editing requires generative models to interpret complex, compositional instructions and modify only the desired elements across modalities, a capability largely beyond existing systems. Current approaches are typically limited to simple, single-attribute edits within one modality, since diverse, high-quality editing pairs are hard to obtain, especially for cross-modal tasks where aligned audiovisual (AV) data is scarce. We present CrossEdit, a unified omni-modal editing model for images, audio, and video that performs AV movie scene edits zero-shot through cross-modal transfer. Our key observation is that complex instruction-following, once learned in any modality, generalizes to others. We exploit the additive nature of acoustic signals to procedurally generate large-scale audio editing pairs with compositional instructions, and show that fine-tuning on this synthetic data and self-supervised AV masked reconstruction, alongside a targeted set of curated cross-modal tasks, induces instruction-following that transfers zero-shot to unseen modality and instruction combinations. To evaluate this capability, we release CrossEditBench, a human-annotated benchmark of AV edits on movie scenes, and propose AV-FES, a metric that jointly scores instruction following and consistency. We show that our proposed techniques improve performance on zero-shot AV movie scene editing, while maintaining or improving performance on lip-synced speech editing and standard image, video, and audio editing benchmarks. Demos are available at https://wanchichen.github.io/crossedit/.

0.25

LiG-DETR: Local-in-Global Reassembly in Latent Space for Aerial Object Detection

cs.CV ·2026-10-07 ·Yupeng Zhang, Fangzhuo Gao, Juntao Cheng, Ziyi Zhao, Liang Wan +1 more
Abstract

Aerial object detection faces substantial scale and density variations. Small objects are easily degraded by downsampling and feature compression, while medium and large objects require sufficient global context. Existing methods mainly follow two paradigms: image slicing provides clearer local evidence but relies on independent crop-level prediction and post-processing, whereas feature- and query-level optimization preserves unified inference but operates on already compressed full-image representations, limiting recovery of fine-grained information. This raises a key question: can aerial detection directly acquire high-fidelity local evidence before feature degradation and integrate it into a unified end-to-end framework? To this end, we propose LiG-DETR, an Efficient Global-Local Reassembly framework that reformulates image slicing as high-fidelity local feature acquisition. A shared encoder extracts global and locally magnified features, which are projected into the detector feature space. The projected local features are reassembled according to their original spatial locations to form a globally aligned local feature level, and a single DETR decoder jointly decodes global and local features. To reduce redundant computation, Context-Preserved Selective Reassembly focuses high-resolution encoding on informative regions while preserving a dense feature layout, and Density-Aware Adaptive Query Allocation adapts the decoder query budget using encoder proposal scores. Experiments show substantial gains on small and medium objects while retaining strong large-object performance, with favorable accuracy--efficiency trade-offs and improved cross-domain generalization. The code will be released.

0.25

Mosquito acoustic identification at unseen recording sites: practical limits of common transfer remedies for vector surveillance

eess.AS cs.LG ·2026-10-07 ·Yucheng Gong, Rui Zhou, Binbin Zeng, Qiang Ren, Hongjin Hui
Abstract · Major revision: reframed from a domain-generalisation study to an acoustic monitoring study. All Perch results re-measured in a single envir…

Passive acoustic monitoring could widen mosquito surveillance, but a classifier trained under one set of recording conditions often fails when moved to a new site, device or enclosure. On a nine-species, five-condition corpus of about 271,000 recordings, the published baseline identifies species at 0.881 balanced accuracy under conditions it has seen and 0.175 under held-out ones. Using that corpus's leave-domain protocol, we asked which transfer strategy is most reliable when a program brings a new site online. Using fixed features from a pre-trained bioacoustic model and a small classifier improved held-out-condition balanced accuracy to 0.305, without updating backbone weights and in minutes of training per site. Four widely used additions did not improve transfer: rebalancing training across recording conditions, fine-tuning the backbone and fusing a weaker second modality each reduced accuracy, and a six-term objective reduced to a single alignment term. Rebalancing reversed direction across four encoders, helping weak representations and harming strong ones. Genus-level aggregation was more accurate than species-level classification (about 0.50 against 0.30), suggesting genus-level screening may be the more realistic operating point here. These results come from one corpus under one protocol, and are dataset-bounded guidance rather than validated general rules.

0.25

Beyond Anonymous Captions: Grounding Character Identity in Video Captioning and Question Answering

cs.CV ·2026-10-07 ·Anas Filali Razzouki, Killian Steunou, Khalil Guetari, Thomas Kling, Mounîm El-Yacoubi +1 more
benchmarkcode/data
Abstract

Linking people's appearance and actions to character identities is essential for understanding video narratives. We present a framework for identity-aware video captioning and person-centric question answering that combines automatic character identification, explicit spatial grounding, and task-specific adaptation. Starting from LSMDC v2 movie clips, our pipeline matches detected faces to actor reference images, tracks characters across frames, and builds inputs with identity-linked bounding boxes. A strong vision-language model generates identity-aware captions and questions, which are manually verified and filtered to create a benchmark of 750 captioned clips and 3,000 person-centric questions. We study five grounding strategies combining textual coordinates with visual face or estimated person boxes across Video-MLLM families at roughly 2B, 4B, and 8B parameters and larger frontier models. Combining visual face boxes with textual coordinates yields the most consistent performance across scales and significantly improves overall performance over coordinates alone. Smaller models tend to over-assign known identities when the queried person is not grounded, while larger models better recognize such UNIDENTIFIED cases. We introduce BAC by LoRA fine-tuning Qwen models at 2B, 4B, and 8B scales on about 32K identity-aware captioned clips. Across all scales, BAC outperforms every other evaluated model family of comparable size. BAC-8B reaches 93.20\% overall QA accuracy, ranking behind only GPT-5.6 Sol among the frontier models evaluated in our study. Overall, explicitly communicating who is where, together with lightweight task-specific adaptation, substantially improves identity-aware video understanding without changing the underlying architecture. We release the benchmark, training data, code, and BAC checkpoints at https://github.com/momentslab/beyond-anonymous-captions.

0.25

Fluid-Gen-Zero: Grounding Pretrained Video Generators in Physics without Training

cs.CV cs.GR ·2026-10-07 ·Hong Huang, Yuqiu Liu, Chenyu You, Daniel Martin, Chuhang Zou +1 more
benchmark
Abstract

We present Fluid-Gen-Zero, a training-free framework for physics-aware fluid-object interaction video generation that decouples physical reasoning from appearance synthesis. Our key insight is to delegate motion dynamics to a physics simulator while preserving the appearance modeling capacity of pretrained video generators. We bridge these two domains through a two-level agentic workflow: generation-time planning, where a vision-language model (VLM) agent interprets intent and the simulation rollout to organize generation clips, and latent-space guidance, which injects simulation signals into denoising through region-aware latent wrapping. This plug-and-play design is compatible with current video foundation models. We further introduce a benchmark for fluid-object interaction video generation. Across Tora (CogVideoX-based), VACE and WanMove (Wan-based), Fluid-Gen-Zero consistently improves simulation alignment, reducing object trajectory error by 26.7%-81.5% and fluid fEPE (fluid flow endpoint error) by 67.9%-84.0%, while largely preserving perceptual quality. In a human preference study, raters favor Fluid-Gen-Zero in 55.1%-74.4% of same-backbone comparisons across three backbones, and in 90.4%-94.2% of comparisons against simulation-based methods. Code and data will be released upon acceptance.

0.25

UltraText Bench: A Comprehensive Bilingual Benchmark for Evaluating Visual Text Rendering in Image Generation

cs.CV cs.AI ·2026-10-07 ·Deyuan Liu, Yihao Hu, Jingxuan Zhang, Xingying Li, Jun Xie +11 more
benchmarkcode/data
Abstract

Dense visual text requires image generators to reproduce long strings across multiple regions with correct placement and legibility. As short-string rendering improves, evaluation must test sustained performance across more demanding scenes. We introduce UltraText Bench, a bilingual benchmark for prompt-only generation of dense visual text. It contains 432 prompts spanning 24 real-world scene categories and three difficulty levels, split equally between English and Chinese. Each human-reviewed prompt supplies exact strings for four to twelve text regions, paired with structured references for their content, placement, and visual attributes. We use the Q-Judger vision-language model to assess each image against the complete reference, reporting text fidelity, text clarity, spatial quality, and scene quality. Across 24 model configurations, these dimensions reveal different strengths: Z-Image-Turbo gains 3.81 clarity points over Z-Image-Base while losing 14.76 fidelity points under the reported settings. Performance also varies with workload; Qwen-Image-2512's English composite falls from 86.50 at L1 to 42.86 at L3. Ten participants took part in human evaluation of the automatic scores. Repository: https://github.com/LINs-lab/UltraText_Bench.

0.25

Training Advisors for LLM Agents from Task Outcomes

cs.CL cs.AIcs.LG ·2026-10-07 ·Sergei Polezhaev, Barys Liskavets, Ori Press, Alexander Golubev
benchmark
Abstract · 26 pages, 11 figures

Large language model agents tackle multi-step tasks by interleaving reasoning and tool calls with observations from the environment. Prior work has shown that natural-language feedback can help these agents revise their decisions during task execution. We introduce Caddie, a method for training critics to provide natural-language analysis and advice as agents work through a task. Unlike approaches that rely on step-level labels or reference critiques, Caddie learns from whether the agent ultimately succeeds after receiving the critic's feedback. We optimize the critic through reinforcement learning while keeping the base model frozen. Trained on multi-hop question answering with a single base model, our Qwen3-4B critic improves success rates across four base models of different scales and architectures, including three not used during critic training. On the MuSiQue benchmark, the trained critic improves Qwen3-4B's success rate by more than 25 percentage points, surpassing the performance of Kimi K3 without a critic. The same critic also yields gains on out-of-domain interactive benchmarks, including $τ^3$ and DeepDive, with no additional training. Our results show that agents can decide when to seek help from a critic at inference time and that outcome-based critic training can produce guidance that transfers across base models and task domains.

0.25

emg2face: Expressive Facial Animation with High-Density Surface EMG

cs.GR cs.CVcs.ETcs.LG ·2026-10-07 ·Ganidhu Abey, Wendy Greening, Ashika Kamboj, Leonhard Helminger, Abhijeet Ghosh +3 more
Abstract · 12 pages plus supplmentary material

Facial movements convey subtle and important information that is critical for human social communication. Optical methods for face capture are difficult or impossible to use when the face is occluded by head-mounted devices (HMDs), such as VR headsets. Even with a clear line of sight, such methods raise privacy concerns and require head-mounted capture rigs that offset cameras and lighting from the face. We show that high-density surface electromyography (HD-sEMG) provides a viable non-optical alternative that addresses these challenges. We measured 64 EMG channels, using two textile EMG grids, with 32 from the forehead (typically occluded by an HMD) and 32 from the side of the face. EMG data were digitized at 2048 Hz and filtered. Facial movements were simultaneously recorded and used to estimate 478 3D facial landmarks using MediaPipe's Face Landmarker. A major challenge in such multimodal recordings is synchronizing EMG and video data, which have different sampling frequencies and independent clocks. We developed a novel synchronization method using analog audio bursts that is capable of sub-millisecond synchronization. We also developed a staged fitting method that fits a recent high-resolution parametric head model (GNM), with 253 identity blendshapes and 383 expression blendshapes, to the MediaPipe landmarks as participants performed different facial expressions. We trained a deep neural network comprising per-grid spatial encoders followed by a dilated temporal convolutional network (TCN) to predict blendshape parameters from HD-sEMG signals at 100 Hz. Once trained, the network can predict expression blendshapes solely from HD-sEMG recordings. The output can be rendered using standard real-time blendshape animation methods. We demonstrate the methods using recordings from 25 participants, and direct expression transfer to a variety of human faces and non-human characters.

0.25

Conversational Voice Aesthetic Model with Reinforcement Learning from Human Listeners

eess.AS cs.CLcs.LGcs.MM+1 ·2026-10-07 ·Xilin Jiang, Shun Zhang, Tejas Jayashankar, Yinghao Aaron Li, Osama Hanna
code/data
Abstract

We introduce Conversational Voice Aesthetic Model, a speech large language model for describing the voice aesthetics of real or synthetic speech responses in natural conversational contexts. Given a context and a response speech, CVAM describes salient moments that characterize the voice and predicts nine categorical attributes spanning gender, pitch, pacing, emotion, and delivery. The key challenge lies in perceptual fields such as emotion and delivery, which are inherently subjective and lack definitive ground truth. Therefore, we collect ~10 human annotations for each of 3k real and synthetic responses derived from the CANDOR corpus. CVAM is supervised finetuned on synthesized aesthetic descriptions and labels, then optimized with Group Relative Policy Optimization on human judgments. Experiments show that CVAM better agrees with human listeners than Gemini 3.1 Pro and open-source speech LLMs, and outperforms single-human-vs.-rest agreement. Together, we demonstrate the importance of grounding voice aesthetics in human perception and propose a principled framework for human alignment.

0.25

Dialect-Robust Speech Language Models with Synthetic Pseudo-Dialect Augmentation

cs.CL eess.AS ·2026-10-07 ·Shunsuke Mitsumori, Tomoya Mizumoto, Yusuke Fujita
published
Abstract · 7 pages, 1 figure, 7 tables. Accepted to IEEE SLT 2026

Speech Language Model (SLM) performance often degrades on dialects due to data scarcity. Conventional text-to-speech (TTS) augmentation struggles to cover diverse dialects as it requires a certain amount of real dialect speech. We propose synthesizing pseudo-dialect speech by converting LLM-generated dialect text via a standard-language TTS model, requiring zero real dialect speech. Additionally, we introduce intermediate standard-text prediction during training, acting as semantic normalization for downstream tasks. We evaluate dialect understanding via dialect-to-English speech translation across Japanese, German, and Chinese dialects. Compared to synthetic standard speech baselines, pseudo-dialect augmentation improves scores for Japanese (from 25.38 to 26.24) and German (from 31.57 to 32.47). Furthermore, the intermediate standard-text prediction effectively bridges the semantic gap, boosting performance to 28.26 for Japanese and from 11.67 to 16.37 for Chinese. These results suggest that our approach scales to various languages without requiring speech resources specific to each dialect.

0.24

Performance at What Cost? A Sustainability-Aware Performance Index for Cell and Nucleus Instance Segmentation

cs.CV cs.AI ·2026-10-07 ·Eiram Mahera Sheikh, Alaa Tharwat, Wolfram Schenck
benchmark
Abstract

Pretrained models for cell and nuclear instance segmentation differ substantially in architecture, pretraining data and objectives, parameter count, inference strategy, adaptation requirements, postprocessing pipeline, and computational demand. Large pretrained and foundation models are increasingly adopted because of their strong zero-shot capabilities, but their use also imposes greater energy consumption, memory requirements, computational demands, adaptation costs, and operational carbon emissions. Whether these additional demands are justified by meaningful gains in segmentation performance remains unclear. We address this question by introducing the Sustainability-Aware Performance Index (SAPI), a configurable metric that combines segmentation performance, energy consumption, and model size. We benchmark 19 pretrained and foundation models across six CellBinDB datasets under zero-shot inference and evaluate 16 fine-tunable models using few-shot adaptation with both frozen encoder and full-model fine-tuning. We estimate energy consumption for GPU, CPU, and RAM using software-based monitoring tools. Our results show that larger and more computationally demanding models do not consistently achieve proportionate improvements in segmentation quality. While few-shot adaptation benefits several models, the gains and resource costs vary considerably across architectures, datasets, and adaptation strategies, causing SAPI-based rankings to differ from rankings based on performance alone. This study provides a practical framework for comparing segmentation models more comprehensively and supports more computationally accessible and environmentally responsible model selection in biomedical image analysis.

0.24

From Pareto to Preference: Personalized Test-Time Scaling via Amortized Agentic Policy Discovery

cs.CL cs.AI ·2026-10-07 ·Xinglin Wang, Zishen Liu, Tong Zheng, Shaoxiong Feng, Peiwen Yuan +7 more
Abstract · Preprint

Test-time scaling (TTS) improves the reasoning capabilities of large language models by allocating additional inference computation. Existing approaches to improving TTS efficiency largely optimize accuracy against one resource dimension at a time, advancing either the accuracy--cost or accuracy--latency Pareto frontier. Yet user requirements are multidimensional: users may specify accuracy, latency, and inference-cost requirements jointly, and different requirements can favor different controllers. We formulate Personalized Test-Time Scaling as discovering executable controllers that maximize the joint satisfaction rate of user-specific requirements. To reduce the overhead of repeated policy discovery for new user profiles, we propose PersonTTS, an amortized agentic policy-discovery framework that reuses prior search experience through requirement-matched controller initialization and source-distilled procedural guidance, while retaining target-profile evaluation for every candidate. Experiments on AIME and HMMT show that PersonTTS substantially outperforms strong TTS baselines in joint requirement satisfaction on unseen user profiles and held-out problems. Under the same candidate-evaluation budget, cross-user experience reuse further improves policy quality while substantially reducing discovery-agent time and cost.

0.24

Event-Aligned Visual Action Reasoning for World Action Models

cs.CV cs.RO ·2026-10-07 ·Xiaomeng Yang, Yushu Wu, Yi Gao, Yuhao Lei, Xuan Zhang +2 more
positioncode/data
Abstract · Project Page: https://xiaomeng-yang.github.io/Event-aligned-WAM/

World-Action Models (WAMs) utilize future visual prediction as an intermediate reasoning process to guide action generation. However, existing WAMs typically structure visual imagination according to predefined temporal intervals, without explicitly accounting for the different roles of task-critical interactions and connecting transitions. We argue that effective visual foresight should align directly with task-relevant interactions and their corresponding reasoning demands. To this end, we introduce an event-aligned visual action reasoning framework that organizes visual-action prediction around interaction events. Through event-aligned visual-action supervision, WAM learns to generate event-aligned visual context in each imagined rollout, placing greater emphasis on critical state changes that inform action generation. This shapes the visual reasoning granularity according to the underlying interaction dynamics, with detailed reasoning around task-critical events and coarser progression through connecting transitions. Furthermore, we introduce an execution validity head that identifies the valid portion of each predicted action sequence, avoiding redundant actions during chunked inference. Experiments demonstrate a 10.26 percentage point improvement in DOMINO success rate over baseline and competitive performance on RoboTwin 2.0. It also transfers from DOMINO Level 1 to Levels 2 and 3 without target-level adaptation.

0.24

LoomSC: Scalable Deep Subspace Clustering with Projector Factorization and Exact Spectral Reduction

cs.CV cs.AI ·2026-10-07 ·Nairouz Mrabah, Youssef Melki, Mohamed Bouguessa, Riadh Ksantini, Shakeeb Murtaza +1 more
Abstract · 19 pages, 7 figures, 5 tables; includes appendices

Dense self-expression matrices and full-affinity spectral clustering limit the scalability of subspace clustering. We introduce the Latent Orthogonal Optimization Model for Subspace Clustering (LoomSC), a framework that addresses both bottlenecks through projector factorization and exact spectral reduction. Motivated by the spectral structure of least-squares regression, LoomSC jointly learns latent features and a projector self-representation through two thin factors. Alternating Procrustes and least-squares updates preserve the sample factor's orthogonality while keeping the coefficient matrix implicit. We construct a nonnegative quadratic affinity that preserves the projector's support. An exact feature map then reduces its normalized spectral problem to an eigenproblem whose dimension depends only on the factor width. Neither the full affinity nor the sample Laplacian needs to be formed. Our analysis quantifies the projector approximation and identifies conditions for subspace preservation and within-subspace connectivity. For fixed dimensions and iteration budgets, the complete pipeline has linear time and memory complexity in the number of samples. Across five image-clustering benchmarks, LoomSC ranks first or second in all 15 dataset-metric comparisons against 9 state-of-the-art baselines. Its mean accuracy exceeds the highest baseline mean by 6.66 percentage points. Synthetic experiments scale to 500,000 samples while maintaining at least 99.8% accuracy.

0.24

Learning to Act with Task Progress: Distilling Small Agents from Compact Teacher Supervision

cs.CL ·2026-10-07 ·Wenxi Gan
Abstract

Learning from large-model demonstrations offers a way to train small agents that can complete recurring tasks without calling a large model at every step. A central design choice is what to retain from teacher trajectories that contain reasoning, actions, and information about task progress. We introduce Task-Progress Distillation (TPD), an offline approach that pairs each demonstrated action with a short label describing the current task stage. The student learns these compact targets and selects actions by jointly scoring admissible stage--action pairs, which a deterministic harness executes in the environment. On ALFWorld, a 1.7B student trained with 404 demonstrations achieves 72.4\% mean unseen task success with either TPD or action-only supervision, compared with 48.3\% for a reasoning-trained student using constrained action selection. Explicit stages provide an additional benefit at 200 demonstrations, improving success from 48.0\% to 67.7\% over action-only supervision. With more demonstrations, the action-only student closes the gap, and both approaches reach 76.9\% at 808 demonstrations. Shared-history analyses link part of TPD's local advantage to better decisions when moving between subgoals, particularly from object acquisition to processing. These results show that compact supervision can train effective small task agents, while explicit task progress provides additional guidance at an intermediate demonstration budget.

0.24

From Single-Shot QA to Continuous Perception: Revisiting the Energy Composition and Optimization Boundary of Edge VLMs

cs.CV cs.AI ·2026-10-07 ·Yiting Duan, Miaoxin Wang, Aoqing Guo, Bo Peng
published
Abstract · 5 pages, 3 figures. Accepted at IEEE International Conference on Data Mining 2026 Teen Research Track

Visual-prefill energy conclusions for edge vision-language models (VLMs) are commonly drawn under single-shot workloads; we revisit them under continuous perception. Profiling nine VLMs across three architecture families on Jetson AGX Thor and Orin (4,508 records, calibrated with a pure-text task), we find two mutually independent energy laws: the per-output-token decode cost grows monotonically with model scale, whereas the visual-prefill (image-encoding) cost is dominated by visual-encoder architectural efficiency, not scale. Under continuous perception ($N{=}10$) with sparse output, visual-prefill energy reaches 92--97\% of total. We derive optimization strategies---visual frame reuse, adaptive output truncation, task-aware routing, and encoder throughput optimization---for edge VLM deployment.

0.24

Reading Position Is the Baseline to Beat: A Time-Ordered Evaluation of Personalised Highlight Prediction

cs.IR cs.HC ·2026-10-07 ·Kazuki Nakayashiki, Keisuke Watanabe
position
Abstract · 13 pages, 1 figure, 5 tables. Ancillary files include the specifications, the results write-ups, the analysis scripts, and the aggregate art…

A reader's first highlights on a page are the cheapest personal signal a reading product has. The natural plan is to suggest what similar earlier readers marked, and to judge the result against popularity. We argue that the baseline to beat is reading position. In a time-ordered evaluation on one social highlighting platform (7,343 reader-page pairs on 1,511 pages after one highlight), ranking the sentences just below a reader's first highlight, with no other reader's data, puts the next highlight in the top five 47% of the time, against 26% for popularity and 29% for the better of two similarity methods. The baseline depends on the target: over all later highlights that ranking loses to popularity, while popularity discounted by distance from the latest highlight, at the scale with the best average precision of three tried, beats popularity and both similarity methods on both targets. In a comparison specified in advance, neither similarity method shows a gain over popularity in average precision over all later highlights, from one to five highlights, and a gain of +0.01 is excluded. Nor would a gain by itself show that a method has found a reader's preferences: synthetic readers who share one set of preferences produce one, and an evaluation out of time order shows a method where the reader went. The position results are exploratory and unconfirmed. Personalisation inside a document should be evaluated in time order and against reading position.

0.24

Do Image Editors Follow Depth-Dependent Blur and Aperture Response? A Rendered-Ground-Truth Pilot Audit

cs.CV ·2026-10-07 ·Zhihan Chen, Yuhuan Zhao, Yijie Zhu, Xinyu Yao, Mengcong Ren +2 more
unusual
Abstract

General image editors are asked to make a photo look as if it were taken at f/1.4, yet it is rarely checked whether the blur they add follows thin-lens optics. A physical aperture edit spreads blur across depth in thin-lens proportions and changes the blur when the aperture changes; prior evaluations check blur monotonicity, sharpness-trend correlation, effective-aperture error, or vision-language judgments, and none we found reports the two properties separately at known depths. In this pilot audit of two editors (Gemini~3.1 Flash Image and GPT-image-2.5) we compare against a rendered oracle: Blender Cycles scenes with true thin-lens depth of field, one blur-width estimator applied identically to oracle and editor outputs, and preregistered depth and aperture indices. In 24 texture scenes rendered in one three-panel geometry, accepted and measurable panels show f/1.4-to-f/2.8 width ratios, $\barσ_{1.4}/\barσ_{2.8}$, of 0.99--1.18 against 1.98--2.13 for the oracle, and ratios of pooled median Gaussian-equivalent near/far blur widths of about 1.18--1.25 (Gemini) and 0.97--1.05 (GPT-image) against 1.69--1.82. Preregistered black-box interventions show that qualitative wording changes blur strength by roughly 2--10 times, whereas a request for 2 versus 6 px changes it 1.1--1.3 times and no tested wording of the f-number meets the registered ``followed'' criterion. The depth compression appears in the original and reversed centre-focus layouts; with the focus on the near panel, the available-panel depth index reaches the registered threshold, and the aperture response stays attenuated in every layout tested. Scalar metrics adapted from published ones give oracle-like scores to synthetic editors whose proportions are compressed.

0.24

Seeing Through the Glare: A Multi-Source Benchmark and Ocular-Adaptive Pixel MeanFlow for Eyeglass Reflection Removal

cs.CV ·2026-10-07 ·Tao Liu, Youwei Pang, Kailai Zhou, Jiaming Zuo, Hanqi Liu +5 more
benchmark
Abstract

Eyeglass reflection removal is important across smartphone imaging, video conferencing, and other face-centric visual applications. The task is challenging because reflections range from mild photometric contamination to severe ocular occlusion, requiring selective correction and plausible reconstruction without altering identity or natural appearance. Existing datasets cover limited reflection conditions, constraining generalization to complex real-world scenes and systematic evaluation. We introduce \textbf{OcuBench}, a multi-source benchmark comprising 10,280 controllable synthetic pairs, 732 real-input pseudo-pairs, and 458 independent real-world test images, supporting both paired evaluation and assessment beyond generated supervision. We further propose \textbf{OcuFlow}, an ocular-adaptive pixel MeanFlow (pMF) framework for efficient, detail-preserving restoration. It combines geometry-adaptive representation with one-step pMF to focus reconstruction on reflection-obscured ocular regions, together with native-resolution frequency-preserving synthesis to retain reliable observed details. Experiments across diverse reflection conditions demonstrate that OcuFlow achieves consistent advantages in reflection removal quality, ocular fidelity, and efficiency. In a blind user study, it receives $67.32\%$ of selections, $6.2\times$ the next-best share. Both the code and dataset will be released.

0.24

MSU Team at the Explainable Deepfake Detection Challenge 2026: Grounded Artifact Evidence for Deepfake Detection

cs.CV ·2026-10-07 ·Artem Filippov, Aleksandr Gushchin, Kirill Koltsov, Dmitriy Vatolin, Anastasia Antsiferova
top venue
Abstract

Recent advances in generative image models have made many manipulated images highly realistic, raising the need for detectors that are not only accurate but also able to provide visual evidence for their decisions. In this paper, we present our solution to the Explainable Deepfake Detection Challenge [2] on the XPlainVerse dataset [1], where systems are required to predict whether an image is real or fake and generate both complex and simple explanations grounded in visible forensic cues. Our method follows a modular detection-and-explanation design. For the real/fake decision, we build a multi-backbone detector that combines several DINOv3 models with Mesorch manipulation-localization features, bringing together pretrained visual representations, DCT-aware cues, and multi-scale forensic information. To inject explanation evidence into the detector, we use a Grounding-DINO-based pseudo-mask generation pipeline that converts local artifact descriptions from training explanations into weak patch- level supervision for an Artifact Evidence Map. We further introduce a local patch-level contrastive objective that separates artifact and authenticity evidence in the detector feature space without requiring paired images or pixel-level manipulation masks. For language output, we use class-conditional Qwen3-VL models to generate complex explanations for fake and real predictions, followed by a GRPO-optimized text simplification model. The proposed methods were trained and evaluated on the challenge subset of XPlainVerse. On the full test split, our submission achieves 0.9349 detection accuracy, 0.5571 explanation score, and a 0.7456 final challenge score.

0.24

Cognitive Thermometers: Machine Learning and Logical Complexity

cs.CL ·2026-10-07 ·Shane Steinert-Threlkeld, Jakub Szymanik
positionunusual
Abstract

How does the human mind represent semantic categories? Why do natural languages favor certain meanings over others? Prior explanations have relied on logical definability and complexity, but these are highly sensitive to the choice of logical language, rendering some design choices unmotivated. In this article, we propose that machine learning provides a somewhat more agnostic approach to measuring semantic complexity. We review emerging evidence that logic and machine learning often yield converging results on relative complexity and its resulting effects in semantic typology. Where they diverge, learning appears to be a better explanation than logical complexity. We argue that treating machine learning models as ``cognitive thermometers'' enables a unified approach to complexity that bridges symbolic logic and connectionist AI.

0.23

Real-Time Obstacle Detection and Distance Estimation on Smartphones for Visually Impaired and Distracted Pedestrians

cs.CV cs.AIcs.PFcs.SE ·2026-10-07 ·Abtin Zandi, Sarvin Nami, Roza Ganjipour, Bardia Sabagh Kermani, Amirreza Tarabkhah +1 more
Abstract · 10 pages, 3 figures, 4 tables

According to the World Health Organization, at least 2.2 billion people worldwide live with visual impairment, including tens of millions with severe vision loss. Traditional aids such as white canes or guide dogs remain costly, limited in capability, and inaccessible to many. At the same time, smartphone distraction in public spaces has been linked to elevated risks of collisions with vehicles, trips, and falls. These dual challenges motivate lightweight, smartphone-based solutions that combine accessibility with low latency. We present a Kotlin-based Android application that uses a YOLOv8 detector, retrained to recognize eighteen common urban obstacle classes, to deliver real-time obstacle alerts for both user groups. The detector was optimized with pruning, quantization, and early-exit branches, and deployed through the Android Neural Networks API (NNAPI), which can dispatch inference to GPU, NPU, or DSP accelerators when available. On a mid-range smartphone, these optimizations reduced average inference latency from about 3.2 s for the off-the-shelf YOLOv8n to 0.39-0.48 s; the recommended configuration (INT8 quantization with early exit) runs in 0.42-0.48 s while retaining a recall of 0.81. The application combines the optimized detector with ARCore-based or monocular focal-length distance estimation, so it works with a single smartphone camera and requires no LiDAR, depth camera, or other specialized sensor. Multimodal feedback, including Text-to-Speech guidance for visually impaired users and vibration alerts for distracted pedestrians, supports use across different contexts.

0.23

Gradient-Based Trajectory Optimisation over Continuous Poses for Sparse-View Cone-Beam CT

cs.CV ·2026-10-07 ·Linda-Sophie Schneider, Simon Wittl, Gabriel Herl, Andreas Maier
Abstract · Submitted to TPAMI

Trajectory optimisation for cone-beam computed tomography (CT) determines which information sparse-view scans acquire. Fixed candidate pools prevent off-grid refinement and require new object-specific precomputation for each acquisition manifold. We make every source pose an individual continuous variable and move all poses jointly by gradient ascent on the scanner's kinematic manifold. The objective combines soft-Tuy plane coverage, continuous View Covariance Loss, and an analytic attenuation-aware ray-bundle penalty. The same optimiser handles circular, limited C-arm, two-axis, and freesphere parametrisations. On a Defrise flange, continuous selection recovers laminar defects invisible to a circular orbit, matches discrete swap search on the free sphere at the sparser budget, and leads at the denser one, with the same objective evaluated in every arm. A moderate elevation band already recovers most of the free-sphere gain at the defects, so the same optimiser transfers to bounded scanner envelopes. Photon noise preserves the ordering on the flange and compresses it on a dense fuel nozzle. Sparseprescan planning benefits from matching prescan and planned acquisition manifolds. Selection takes seconds rather than minutes without an object-specific reconstruction basis. Prescan-planned poses were executed on a robot CT bench and reconstructed in a common frame, demonstrating feasibility but no consistent metric gain over uniform band sampling. Continuous pose optimisation incorporates attenuation and scanner constraints directly into sparse-view acquisition design.

0.23

From Global Alignment to Local Grounding: Zero-Shot Chinese Character Recognition with Radical Verification

cs.CV ·2026-10-07 ·Yu-Heng Shih, Bing-Chen Wu, Tsz-To Wong, Ting-En Yen, Hong-Han Shuai +4 more
benchmark
Abstract

Zero-shot Chinese character recognition (ZS-CCR) aims to recognize characters whose categories are never observed during training, and typically relies on the compositional structure shared between seen and unseen characters. Recent CLIP-style methods represent this structure with the Ideographic Description Sequence (IDS) and align it with glyph images in a shared embedding space. However, they rely on a single global image--IDS similarity that discards the spatial layout of radicals and, being learned only implicitly from seen classes, generalizes poorly to unseen ones; moreover, global matching often retrieves the correct character within the top candidates yet fails to rank it first when characters differ only in subtle local radicals. To address these issues, we propose a global-to-local two-stage framework. In the first stage, STG-CLIP augments the IDS with explicit tree-position and radical-level geometric priors, yielding a spatial-aware prototype that provides a consistent spatial description across seen and unseen categories for high-recall global retrieval. In the second stage, the Radical Verification Module (RVM) uses the radical instances of each retrieved candidate as queries to verify whether the corresponding radicals can be matched to spatially compatible regions in the input glyph. A margin-based gating rule activates the RVM only when the leading global candidates receive similar similarity scores. Experiments on the ICDAR2013 benchmark demonstrate that our method achieves state-of-the-art performance under the character-level zero-shot setting, obtaining 83.06% top-1 accuracy with 2,755 seen classes. Ablation studies further show that the explicit geometric priors and radical-level verification provide complementary improvements.

0.23

The Impact of Backbone Evolution on LLM-Based Relevance Assessments

cs.IR ·2026-10-07 ·Chuting Yu, Guido Zuccon, Teerapong Leelanupab
Abstract · 12 pages main content

LLMs are evolving rapidly, with newer models offering stronger capabilities. This suggests that in LLM-based relevance judging, more capable models will achieve higher agreement with human judgements under the same prompt. We challenge this understanding by investigating the behavior of LLM-based relevance judges under backbone evolution. Keeping the prompts fixed, we evaluate a representative single-prompt (UMBRELA) and a rubric-based prompt (EXAM) across sequential model versions of commercial (Gemini, GPT) and open-weight (Qwen, Llama) models. Overall, we find no consistent evidence that newer versions lead to better relevance judges. Crucially, similar or improved aggregate performance does not imply judgment stability: correct judgements made by an earlier version of an LLM backbone are not necessarily preserved by later versions. We investigate the potential drivers of these regressions. Our findings caution against the assumption that judging prompts designed and validated for one backbone version will perform equivalently or better when the model is updated, even within the same family.

0.23

CRT-HMAR: Causal Requirement Tracing-Guided Hierarchical Multi-Agent Regulation for Open-Task-Aware Infrared-Visible Image Fusion

cs.CV ·2026-10-07 ·Zengyi Yang, Shuai Yuan, Zhong-Cheng Wu, Juan Cheng, Huafeng Li +1 more
Abstract · 17 pages, 11 figures

Infrared and visible (IR-VIS) image fusion integrates complementary multimodal information into a single fused image to support downstream vision tasks. However, existing methods are typically tailored to seen tasks within a fixed task set and struggle to generalize to unseen tasks, which restricts their applicability in real-world open-task scenarios. To address this issue, this paper proposes CRT-HMAR, a Causal Requirement Tracing-Guided Hierarchical Multi-Agent Regulation Framework for open-task-aware IR-VIS image fusion. CRT-HMAR introduces a Causal Requirement Tracing Task Localization mechanism, which actively intervenes in key image information and observes task-network response variations to map task-specific semantic preferences into image-level causal requirement maps. Based on these maps, a requirement analysis agent aggregates task-specific requirement knowledge to adaptively guide requirement-customized image fusion. Moreover, CRT-HMAR incorporates History-Analysis Multi-Objective Balancing and Task-Level-Correction Conflict Mitigation mechanisms, jointly constructing a hierarchical regulation chain of "requirement interpretation - task balancing - conflict mitigation". Through multiple collaborative agents, CRT-HMAR dynamically regulates key processes including open-task requirement modeling, multi-task balanced optimization, and gradient conflict mitigation. Extensive experiments on open-task scenarios involving five downstream tasks demonstrate that CRT-HMAR significantly improves generalization to unseen tasks while maintaining the performance and balance of seen tasks. Overall, CRT-HMAR shifts IR-VIS image fusion from task-oriented modeling toward requirement-oriented modeling, promoting its extension from closed-task settings to real-world open-task scenarios.

0.23

AbridgeKV: Compressing Long Reasoning into the KV Cache of Learned Pause Tokens

cs.CL cs.LG ·2026-10-07 ·Jack Cai, Lijie Yang, Vasilis Kontonis, Dimitris Papailiopoulos, Shivam Garg
Abstract

Long chain-of-thought decoding is memory-bound: the KV cache of the generated tokens grows linearly with the test-time compute being spent. Evicting cache entries does not fix this, because a model never trained on a cache with entries removed generates longer, often up to the length limit. We instead train the model to compress its own output. AbridgeKV compresses each completed block of generated text into the KV entries of a few learned pause tokens. These are extra input tokens with trained embeddings that produce no output. Once the pause tokens have encoded the block, AbridgeKV frees the block's original cache; subsequent tokens attend to the pause tokens' KV entries in its place. Three novel choices make this work: the pause tokens' KV entries take positions inside the block they replace, the cache budget is allocated across layers according to the measured value of each layer's entries, and compression operates on relatively large blocks. On 446 competition problems from AIME, HMMT, AMC, BRUMO, CMIMC and SMT with 32K-token generations, keeping a quarter of the generated cache matches an identically trained full-KV model in accuracy and output length; the strongest published evictor at a matched budget writes far longer, and the closest trained compressor, retrained on the same data, loses 10 points. Served in vLLM, the cache is 3.5x smaller per sequence and a saturated H100 finishes a problem 3.2x faster than with full KV.

0.23

ScribbleEdit: A Benchmark for Scribble-Only Image Editing

cs.GR cs.CV ·2026-10-07 ·Jie Ren, Hao Kang, Kai Guo, Yiding Yang, Bo Liu +7 more
benchmark
Abstract

Scribble-based interaction provides a lightweight and intuitive way for users to specify image editing intents in interactive editing tools. However, current image editing models based on VLMs or LLMs struggle to understand and execute edits based solely on scribble inputs. To systematically study this problem, we construct a new benchmark, ScribbleEdit, that evaluates the ability of image editing models to perform image editing conditioned on scribbles. This task requires both a deep understanding of the intention of the scribble and an accurate interpretation of its spatial information. In ScribbleEdit, we design an automated data construction pipeline and introduce a dedicated evaluation protocol that explicitly measures intention alignment. Our analysis reveals that existing VLM/LLM-based editing models fail to accurately capture scribble intentions. To guide future progress on scribble-only image editing, we propose a simple yet effective soft-token baseline, which enhances the model's understanding of scribble semantics and outperforms standard image editing models on our benchmark. Our evaluation and baseline together provide a concrete foundation for assessing and improving the scribble-driven image editing.

0.23

Back in Style: A Sociolinguistic Approach to Authoring and Measuring Persona Fidelity in User Simulation

cs.CL ·2026-10-07 ·Lex Konnelly, Elena Khasanova, Riqiang Wang, Matthias Lee, Harsh Saini +1 more
publishedposition
Abstract · Accepted to the UserSim @ NeurIPS 2026 workshop (non-archival)

As agentic systems gain commercial popularity, user simulators increasingly serve as measurement instrument for their evaluation. However, the fidelity of simulated users in comparison to real human users is generally low, and typically assessed by costly, subjective LLM judges. In this pilot study, we ask whether fidelity can instead be measured deterministically by treating a user persona sociolinguistically: as a social type that emerges from observable linguistic style, rather than one predicted by labels or descriptions a model must extrapolate into behaviour. We author personas as concrete stylistic rates, which lets us transfer two established, model-free instruments -- authorship-verification stylometry and lexicon-based content analysis -- as fidelity diagnostics. We A/B-test the sociolinguistic schema against a flat descriptive baseline across five task-oriented customer-service agents. Results show that the sociolinguistic schema improves both stylistic adherence and stylometric distinguishability for most of the tested models, with a caveat that persona style fidelity does not necessarily equal persona "naturalness". We argue that a sociolinguistic approach to persona design is a promising path towards more diverse and representative user personas, and that these metrics are most valuable in an error-attribution analysis, localizing where fidelity breaks down. This is a first step towards interventions that move user simulations closer to faithful renderings of diverse and variable linguistic outputs.

0.23

LinSlot: Exploiting Linear Representation hypothesis for unsupervised attribute discovery from slot based object representation

cs.CV cs.AI ·2026-10-07 ·Sanket Gandhi, Utkarsh Giri, Varun Subramaniam, Rohan Paul, Parag Singla
theory
Abstract

This paper studies the problem of learning disentangled representations of objects and their attributes from raw, unstructured image data. Slot-based methods have shown considerable success in unsupervised learning of object representations from images. Block-slot attention-based methods extend this framework to attribute representations by assuming a uniform factorization of object representations into attributes, which may be suboptimal and consequently limit the quality of the learned representations. We therefore investigate a framework for jointly discovering object and attribute representations. Our key contribution is leveraging the Linear Representation Hypothesis (LRH), which postulates that composable concepts can be represented as linearly additive subspaces in slot representations. Based on this insight, we propose a probabilistic model connecting images, slots (objects), and blocks (attributes). We present an architecture that leverages block attention to connect attribute representations to slots and incorporates LRH in both object and attribute representation spaces. This architecture effectively optimizes the Evidence Lower Bound (ELBO) of the proposed graphical model. Our experiments demonstrate (i) effective discovery of disentangled object and attribute representations, (ii) empirical evidence for LRH in slot space, and (iii) the ability to perform image editing owing to the disentangled and interpretable nature of the learned representations. Our experiments on multiple datasets demonstrate improvements in DCI scores over state-of-the-art methods.

0.23

What Makes Synthetic Hard Negatives Work in Vision-Language Pretraining?

cs.CV cs.AI ·2026-10-07 ·Nikos Giakoumoglou, Paschalis Giakoumoglou, Andreas Floros, Kleanthis Marios Papadopoulos, Tania Stathaki
Abstract · ACCV 2026

Synthetic hard negatives generated in the representation space have proven effective for unimodal self-supervised learning, but transferring this idea to vision-language pretraining is not straightforward. We analyze six representation-space synthesis strategies and identify two failure modes in their transfer to vision-language pretraining: cross-modal constructions that produce overly easy negatives or pull them toward the query, and intra-modal constructions that incorporate the matched positive. We also observe logit-scale saturation when training with synthetic hard negatives and a learnable temperature, and find that fixing the temperature improves downstream performance. Using this geometric analysis we propose SNAP, which generates intra-modal hard negatives that never involve the positive from either modality, avoiding both failure modes entirely. SNAP is model-agnostic, requires no external generative models, and adds less than 10% training time overhead. Evaluated on top of CLIP and FLIP across multiple architectures and datasets, SNAP delivers consistent improvements on zero-shot retrieval, zero-shot classification, and linear probe evaluation.

0.23

Efficient 3D Gaussian Head Avatars for Edge Devices

cs.CV ·2026-10-07 ·Umar Farooq, Jean-Yves Guillemaut, Adrian Hilton, Marco Volino
Abstract

Generative 3D Gaussian head avatars provide high-quality, efficient rendering, but synthesising the Gaussian representation remains computationally expensive, limiting deployment on resource-constrained and edge devices. We introduce an efficient generator architecture for unconditional 3D Gaussian head synthesis, based on a parameter-efficient synthesis block and depth-wise separable convolutions while retaining style-based conditioning. Our architecture reduces generator complexity without requiring model compression or quantisation. Compared with the baseline model, our approach reduces FLOPs by 94%, parameter count by 70%, and model size by 81%, while maintaining competitive generation quality. We further demonstrate practical CPU inference and browser-based execution on mobile devices using ONNX Runtime, enabling 3D Gaussian avatar synthesis without dedicated GPU hardware or application-specific software. In addition to conventional image-quality metrics, we evaluate multi-view consistency, training cost, and deployment performance. Code, trained models, and evaluation tools will be released publicly.

0.23

Velocity Scaling in Flow Matching

cs.CV ·2026-10-07 ·Youssef Saied, François Fleuret
Abstract · 41 pages, including appendix

Scaling a learned flow-matching velocity field $v_θ$ by a gain $γ(t)$ was recently shown to greatly improve generation quality. Prior work argued that velocity fields trained with mean-squared error (MSE) systematically underestimate velocity magnitude and that scaling corrects this error. We show that MSE training does not create a velocity-magnitude deficit. We find instead that velocity scaling reduces population time lag: sampled states at model time $t$ resemble training states from an earlier time. Velocity scaling and moving model time back are two ways to address this population time lag. Across architectures and model sizes, measuring population time lag and using it to select a gain greatly improves generation quality, reducing FID from 28.0 to 12.2 (estimated by linear interpolation between FID measurements at neighboring gains) on ImageNet-256 at NFE 25 without guidance.

0.23

Tropical Cyclone Wind Field Diagnosis with Image-to-Image Models

cs.CV physics.ao-ph ·2026-10-07 ·Sarah J. Ollier, David S. Nolan, Brian D. McNoldy, Eduardo Siman
Abstract · Preprint prior to peer review. Submitted to Artificial Intelligence for the Earth Systems on 8th September 2026

This study presents a novel application of image-to-image generative adversarial networks (GANs) to produce a complete tropical cyclone (TC) surface wind field diagnosis from a single frame of satellite infrared brightness temperature (IR). Because TC surface wind fields are infrequently observed, we train on two data sets. The first contains 3,323 image pairs of IR and wind fields from Weather Research and Forecasting Model (WRF) simulations (WRF2WRF). The second contains 11,886 real IR satellite images paired with parametric wind fields built on the extended best-track data set (IBTrACS) of TC intensities and wind radii (GOES2PARA). Both models, conditioned on storm intensity, motion, and latitude, transform single-channel IR into wind fields over 6$^{\circ} \times$ 6$^{\circ}$ (WRF2WRF) and 7$^{\circ} \times$ 7$^{\circ}$ (GOES2PARA) domains, at 256 $\times$ 256 pixel resolution. GOES2PARA captures outer wind field asymmetries, enabling diagnosis of 64, 50, and 34-knot wind radii (R64, R50, R34) in each quadrant. WRF2WRF also produces detailed rainbands and eyewall asymmetries. Pixel-wise RMSE against held out truth demonstrates realistic wind field generation with mean RMSE of 6.4 knots (GOES2PARA) and 7.2 knots (WRF2WRF). Compared against IBTrACS, WRF2WRF achieved mean absolute errors (MAE) of 14.8 nautical miles (NM) for radius of maximum wind (RMW) and 25.9 NM for R34. GOES2PARA achieved 10.0 NM MAE for RMW and 25.9 NM for R34. Against Synthetic Aperture Radar wind fields GOES2PARA yielded a lower mean pixel-wise error (RMSE 12.1 kt, MAE 9.5 kt) than WRF2WRF (RMSE 13.9 kt, MAE 11.3 kt). Unlike many existing operational surface wind speed algorithms, our image-to-image models provide updated wind fields with every new satellite image.

0.22

VM-ARRAYDPS: Virtual Microphone Augmented Diffusion Posterior Sampling for Unsupervised Blind Speech Separation

eess.AS cs.LGcs.MMcs.SD+1 ·2026-10-07 ·Jingqi Sun, Haozhan Tang, Shulin He, Zhong-Qiu Wang
Abstract · Submitted to ICASSP 2027 and currently under review

Blind Source Separation(BSS) is a fundamental problem in signal processing, aiming to separate multiple source signals from their mixtures without prior knowledge of the sources or the mixing process. Traditional approaches, such as Independent Vector Analysis (IVA) exploits statistical independence of sources. Recently, diffusion-based approaches have emerged as a promising alternative by leveraging powerful generative priors. Among them, ArrayDPS formulates BSS problem as a posterior sampling problem, and utilizes a pretrained speech diffusion model to guide the recovery of clean source signals. A key factor behind its separation capability is the multi-channel consistency (MC) objective, which enforces the estimated source signals to reconstruct the observed microphone mixtures through the estimated acoustic transfer functions. However, the number of microphones in the array is often limited, which constrains the performance of ArrayDPS. To address this issue, we propose VM-ArrayDPS, a novel method that augments the microphone array with virtual microphones with higher-SNR, these microphones can offer extra MC constraints to enhance the separation performance. Experimental results demonstrate that VM-ArrayDPS significantly outperforms ArrayDPS on both 2-speaker and 3-speaker datasets, showcasing the effectiveness of virtual microphone augmentation in improving BSS performance. We also did ablation studies to show the influence of the number of virtual microphones and weight of the MC objective brought by virtual microphones.

0.22

DeepTopoClustering: Unsupervised Derivation of Surface Process Taxonomy from 4D Point Clouds for Topographic Monitoring

cs.CV cs.LG ·2026-10-07 ·Jiapan Wang, Daan Hulskemper, Mathilde Letard, Roderik Lindenbergh, Katharina Anders
Abstract · Submitted to ISPRS Journal of Photogrammetry and Remote Sensing

4D point clouds acquired by permanent laser scanning (PLS) enable accurate high-frequency monitoring of surface change in dynamic topographic environments. However, existing methods remain limited in organizing detected surface activities into meaningful process types. We propose DeepTopoClustering (DTC), an unsupervised framework for deriving a hierarchical process taxonomy from object-based surface activities, so-called 4D objects-by-change (4D-OBCs). We transform each 4D-OBC into a GeoMorphogram, a distributional sequence representing the temporal evolution of topographic change within a spatially bounded surface activity. A convolutional autoencoder learns latent embeddings from GeoMorphograms, which are jointly optimized using a hierarchical deep clustering objective to organize surface activities into a hierarchy. We evaluate the learned hierarchy using expert annotations on two 4D datasets of sandy beach sites and their combination. DTC with GeoMorphograms achieves the highest agreement with expert judgment at the taxonomy level comprising eight major process types ($F_1=0.78$, match accuracy $=0.92$), outperforming dimensionality reduction and conventional flat clustering. The learned taxonomy separates major erosion- and deposition-dominated activities and distinguishes finer subtypes based on change magnitude, duration, compactness, and temporal evolution. DTC thus provides a scalable and interpretable route from 4D change detection to a data-driven, expert-supported surface process taxonomy, advancing automated knowledge derivation for understanding surface dynamics in topographic monitoring.

0.22

MIRROR: From Imitation to Internalization in LLM Personalization

cs.CL cs.AI ·2026-10-07 ·Huayi Lai, Jicheng Yang, Min Yi, Chong Meng
Abstract · 36 pages

The demand for personalized LLMs is shifting from style imitation toward content quality. We investigate whether self-distillation can bridge this gap in existing fine-tuning paradigm. To address this limitation, we introduce MIRROR(Meta- personalization by Internalizing Reference-Revealed On-policy Reflections), a novel self-distillation framework that shifts LLM personalization from imitation toward preference internalization. First, we replace reference-token imitation with reference-revealed on-policy self-distillation, aligning the model's next-token distributions along its own generation trajectories with those of its reference-conditioned self, thereby internalizing user preferences rather than reproducing reference wording.Second, we introduce MIRROR-F, a focal plug-in that augments on-policy distributional alignment with selective supervision over informative reference tokens, thereby strengthening content generation while preserving user-specific expression. Across three personalized generation benchmarks, two model scales, and complementary reference-based and LLM-based evaluations, MIRROR and MIRROR-F achieve leading overall personalization performance and superior text quality, while exhibiting less catastrophic forgetting than SFT-based baselines on three unseen personalized generation tasks. The gains are consistent across model scales and application scenarios, translating to improved performance in LLM personalization tasks.

0.22

PCAsplat: Gaussian Splatting with Local PCA Regularization

cs.CV cs.GR ·2026-10-07 ·Vitor Matias, Filipe Nascimento, Kiyohiro Nakayama, João Paulo Lima, Márcus Lobo +4 more
Abstract

Gaussian splatting has emerged as a flexible representation for 3D reconstruction from posed images. However, existing methods are optimized primarily using rasterization-based losses, which supervise a splat only when it contributes to sampled camera rays. Gaussians that are occluded or contribute little to the sampled view therefore receive weak or no geometric gradients and may drift away from the underlying surface, producing undesired floaters. We introduce PCAsplat, a geometry-aware regularization framework for Gaussian splatting based on differentiable local principal component analysis (PCA). Our PCA regularizer acts directly on neighborhoods of Gaussian centers and can therefore update Gaussians that do not contribute to the current training view. We regularize the PCA eigenvalues to encourage Gaussians to move to the underlying surface with isotropic tangent-plane coverage. We also align each Gaussian normal with the PCA-estimated neighborhood normal to enforce consistent orientation. Experiments on DTU, Tanks and Temples, and NeRF Synthetic show that the splats produced by PCAsplat better approximate samples of the reference surface while substantially reducing undesired floaters. These surface-aligned splats enable downstream geometry-processing tasks, including point cloud segmentation, and direct Poisson reconstruction. Additionally, PCAsplat remains competitive under conventional novel view synthesis and mesh extraction tasks. Code will be released.

0.22

A Study on Improving Multi-class Audio Source Separation Via Decoupled CLAP Query Optimization and an Automated Data Engine

cs.SD ·2026-10-07 ·Amirhossein Hajavi, Hanhee Lee, Pushya Jain, Sky Qiao, Emmanuel Ko +2 more
code/data
Abstract · Project page: https://amhajavi.github.io/AudioSourceSeparation

Language-queried audio source separation (LASS) enables extracting any sound source using natural language. However, adapting LASS models to application-specific sound classes is challenging due to noisy training data and limited semantic coverage of the CLAP-based control signals. We propose a framework comprised of an automated data engine for training-data curation and a two-stage optimization process for class-specific CLAP control signals. Our objective evaluations across seven sound classes show that data refinement and control signal optimization consistently improve source separation performance. Subjective evaluation with 17 participants further demonstrates perceptual improvements of the model trained with optimized control signals over baseline and similar commercial models.

0.22

Relational Abstractions for Spatial Reasoning with Diffusion Models

cs.CV ·2026-10-07 ·Ana Ezquerro, Ozan Özdenizci
benchmark
Abstract

Diffusion models excel at image synthesis, but they remain limited in their ability to reliably satisfy structured spatial reasoning constraints. In conditional data distribution modeling tasks with implicit logical structure, such as puzzles defined by visible clues paired with consistent solutions, state-of-the-art generative models tend to approximate pixel-space distributions without learning the underlying logical rules required for inference. To address this limitation, we present a novel framework for spatial reasoning with diffusion models that leverages unsupervised object discovery and abstractions of object relations. We show that the relational knowledge derived from object-centric representations enriches diffusion models with structural primitives, allowing them to effectively guide the generative representation space during both training and inference, and enabling conditional image generation that satisfies reasoning constraints. Additionally, we introduce a large-scale generative spatial reasoning benchmark with four datasets inspired by human-solvable puzzles. Our results show that relational abstractions significantly improve reasoning capabilities of diffusion models on a variety of complex reasoning tasks, while enabling robust generalization in out-of-distribution settings.

0.22

HarnessIR: Harnessing Multimodal Foundation Models for Universal Real-World Image Restoration

cs.CV ·2026-10-07 ·Xiangtao Kong, Shuaizheng Liu, Rongyuan Wu, Lingchen Sun, Zhengqiang Zhang +3 more
benchmarkcode/data
Abstract

Real-world low-quality images suffer from complex mixed degradations, including but not limited to noise, blur, atmospheric effects, etc. Recent agentic methods usually model real-world image restoration (Real-IR) as a sequential tool calling problem over task-specific single-degradation restoration models. This paradigm, however, is fundamentally limited because complex real-world degradations cannot be cleanly undone degradation by degradation, and the tool used for task-specific models caps the capability of the agent system. In this work, we present HarnessIR, an agentic framework for Real-IR by harnessing a multimodal foundation model (MFM) as the executor. HarnessIR consists of five stages: perception and diagnosis, on-demand tool invocation, prompt composition, execution, and verification-driven refinement. Unlike prior agentic Real-IR methods that rely on tool chains assembled from task-specific models, HarnessIR feeds the restoration requirements, the perceptual diagnosis, and the evidence into an MFM that performs restoration in a single pass, followed by verification stages to determine whether the result warrants further processing. Under our harness, off-the-shelf MFMs handle restoration tasks remarkably well, achieving state-of-the-art results on the widely used MiO100 synthetic benchmark. More importantly, by exploiting the strong generalization ability of MFMs, HarnessIR delivers compelling restoration quality on challenging real-world scenes where previous agentic IR systems often struggle. Codes is available at https://github.com/PolyU-VCLab/HarnessIR.

0.22

Cache the Encoder Within:Compact, Reusable Memory across LLM Queries

cs.CL ·2026-10-07 ·Hanzuo Liu, Chunyu Liu, Chaofan Lin, Alex Lamb, Mingyu Gao
benchmark
Abstract · 17 pages, 3 figures, 7 tables

Repeated queries over shared documents incur redundant encoding, while caching model states introduces persistent storage costs. Building on CoMem's intermediate-state interface, EncBank treats a pretrained LLM's lower layers as a reusable document encoder and compactly stores their outputs for an adapted upper-layer reader. A self-distilled suffix adapter is shared across storage precisions within each backbone, without quantization-specific retraining. Across five benchmark suites on three Qwen backbones spanning different sizes and full-attention and hybrid architectures, 4-bit storage keeps each reported benchmark aggregate within one score point of native-precision EncBank. In a fixed Qwen3-8B workload, it retains 28.1% of the native-precision persistent GPU store. Separate native-precision controls yield a 1.40x selected-pack prefill speedup over same-evidence, same-adapter text replay, at a 3.12-point RULER accuracy cost. A native-precision Qwen3.8-27B configuration also passes 70 of 89 Terminal-Bench 2.1 tasks. EncBank thus combines reusable computation with compact memory, while task fidelity and end-to-end benefits remain dependent on the workload, preparation costs, and reuse frequency.

0.22

Large Language Models for Machine Translation Quality Annotation: Humans and Models Are Both Challenged

cs.CL ·2026-10-07 ·Hala Almaghout, Christian Federmann, Qin Gao
top venue
Abstract · Accepted at EMNLP 2026

Large Language Models (LLMs) are considered to be a more efficient and cost-effective alternative to human judgment for Machine Translation (MT) evaluation. With MT evaluation spanning a large number of language pairs, domains and levels of annotation granularity, LLMs must be thoroughly evaluated across these dimensions before being reliably used as alternatives to human evaluation. In this paper, we evaluate the performance of LLMs for two prominent MT quality evaluation schemes: Multidimensional Quality Metrics (MQM) and Error Span Annotation (ESA) by comparing their agreement with human annotators. We present results on a long-context test set of 70 language pairs and the publicly available WMT23 and WMT25 data, investigating both score and error span annotation agreement across a variety of language pairs and domains. Our results show that while LLM agreement with human annotators exceeds agreement between human annotators for some evaluation tasks, both vary substantially across annotation schemes, language pairs and domains and remain unreliable for most settings. Furthermore, we identify challenges facing both human and LLM annotators: humans are particularly challenged by fine-grained MQM annotations and low-resource language pairs, while LLMs struggle with minor errors, wrong language variants and error span annotation. Our results highlight the potential for improvement for both human and LLM annotation performance, possibly through human-LLM collaborative annotation pipelines that address the reliability issues identified in this work.

0.22

Latent Watermarks under Generative Editing: A Benchmark and Analysis of Detection Survival

cs.CV ·2026-10-07 ·Sung Ju Lee, Nam Ik Cho
benchmark
Abstract

Ordinary prompt-based editing can cause latent watermark detection to fail without explicitly targeting the watermark. We benchmark eight watermark methods against five editors across four generative backbones, four editing strengths, and five semantic categories, with edit-validity and threshold checks. Separating editing from seven subsequent distortions reveals that editing alone primarily distinguishes Tree-Ring, while added distortions expose a broader spectrum of detection survival. Sequential edits reveal a second hidden difference: score separation can decline while detection rates remain near their ceiling. Across methods, standardized clean score separation ($d'$) organizes composite-survival tiers, whereas spatial overlap adds little to predicting edit-only survival beyond clean detectability. Embedding-strength interventions in two methods link higher clean separation to higher post-edit separation. In HSTR, the margin contrast is positive, while the angular layout contrast at matched clean separation remains unresolved. Together, outcome decomposition and continuous separation expose differences hidden by aggregate TPR. Method tiers are stable under threshold recalibration at the main operating points and alternative composite weights. Clean $d'$ is thus a useful empirical diagnostic within this benchmark, with mixed transfer to unseen methods. Code and supporting artifacts are planned for a separate release.

0.22

SAREO-FM: Decoupled Semantic Supervision for SAR-EO Foundation Models

cs.CV ·2026-10-07 ·Jeonghyeok Do, Munchurl Kim
code/data
Abstract · Please visit our project page at https://kaist-viclab.github.io/SAREO-FM_site/

Synthetic aperture radar (SAR) and electro-optical (EO) imagery provide complementary observations: SAR enables day-and-night, weather-resilient sensing, whereas EO provides rich appearance and fine-grained semantic cues. We introduce SAREO-FM, which avoids forcing a single token stream to serve two distinct roles: modality tokens preserve how each sensor observes the scene through masked reconstruction, while learnable semantic queries capture what the scene contains under guidance from a pretrained vision foundation model (VFM). By jointly encoding these queries with SAR and EO tokens, the queries acquire modality-grounded semantic context, while the modality-token outputs remain the explicit targets of masked reconstruction. This design assigns semantic and reconstruction supervision to separate token streams while preserving their interaction within the shared encoder. Pretrained on the million-scale SAR-1M corpus, SAREO-FM achieves strong unimodal transfer for both SAR-only and EO-only inputs, while delivering substantial gains from joint SAR--EO observations on tasks that benefit from complementary sensing.

0.22

Loose Lips: Understanding Privacy Leakage in Agentic Dialogue

cs.CL cs.AI ·2026-10-07 ·Soumya Sharma, Ayana Hussain, Heber Hwang Arcolezi, Ulrich Aïvodji, Spandana Gella +2 more
benchmark
Abstract · Spotlight @ HAIPS workshop @ COLM 2026

As LLM agents are increasingly deployed in high-stakes enterprise settings, understanding how they handle sensitive information in agentic dialogue has become critical. Privacy leakage in such settings requires reasoning about whether information flows respect the contextual norms under which they originated. We extend CIMemories to CIMEMORIES DIALOGUE, a benchmark of 4,410 two-agent dialogue instances grounded in Contextual Integrity theory. Using this benchmark, we systematically evaluate 7 open-weight models from the Qwen3 and Llama 3.1 families across conversation lengths of 1 and 6 turns under three prompting conditions. We find that task capability and privacy leakage are fundamentally coupled: more capable models are systematically leakier, with no model in our evaluation able to decouple the two. Thinking models consistently leak more than their non-thinking counterparts, a pattern we attribute to the extended reasoning surface created by chain-of-thought. Leakage compounds with conversation length, rising up to 3x from 1 to 6 turns. Our results suggest that the coupling between task capability and privacy leakage is not an artifact of misconfiguration but a structural feature of current agentic systems.

0.22

From Digital Human Interactions to Physics-Based Humanoid Skills: Physics-Grounded Post-Training of Interaction Generators

cs.CV ·2026-10-07 ·Kerui Chen, Jianrong Zhang, Kai Lv, Hehe Fan
Abstract

Recent methods have made promising progress in generating interactions between two humanoids, largely relying on physics-based tracking policies to convert digital reference motions into executable trajectories. However, limited tracking capabilities restrict the range of reference motions that can be successfully executed, reducing data utilization. Moreover, even successful tracking does not guarantee physically plausible responses or faithful realization of the intended interactions. In this paper, we introduce DIGHT, a co-adaptive framework that couples a Digital human Interaction Generator with a Humanoid Tracking policy. Our DIGHT first executes multiple text-conditioned interaction candidates in simulation using a fixed tracker. It then constructs physics-grounded preferences from the resulting rollouts, covering both general executability and interaction fidelity. Rather than collapsing these signals into a single scalar reward for candidate ranking, we align the pretrained generator using physics-decoupled diffusion direct preference optimization (DPO), preserving criterion-specific supervision without differentiating through the simulator. To improve executability, preference pairs are derived from tracking error, friction, and floating. Additionally, to improve interaction fidelity, we propose to incorporate force feedback from simulator as a measure of contact fidelity and construct preferences over contact occurrence, location, duration, and force magnitude. The aligned generator then supplies reference motions for fine-tuning the tracker, improving compatibility between generation and physical execution. Extensive experiments demonstrate that our approach not only improves the physical plausibility of generated motions but also enables more reliable and faithful humanoid interactions in simulation.

0.21

Visual Jev Rewards: Reference-Bound Verification for Multi-Subject Image Generation

cs.CV cs.AIcs.LG ·2026-10-07 ·Baoteng Li, Wenzhuo Wu, Kongming Liang, Zhanyu Ma
Abstract

Multi-subject image generation requires rewards that verify whether requested attributes, actions, and relations hold for the specified reference subjects. Subject presence alone does not establish that the correct subjects participate in a requested interaction. We present reference-bound Visual Jev rewards that turn these visual decisions into generator training signals. Each subject-related question receives a positive label only when the requested condition and the relevant reference identities hold jointly. We construct fixed questions offline, train a Qwen3.5-4B verifier with binary supervision, and directly read Yes probabilities from its language-model head. Their mean supplies a GRPO reward while retaining individual judgments for inspection. Using 200 MICo-150K training tasks and 30 updates, the framework raises a GPT-5.4 composite score from 41.78 to 52.50 on a manually selected 897-task MICo-Bench subset; direct 27B rewards yield 51.84. Each reward is tested in one GRPO run, and offline human evaluation does not establish a statistically significant advantage over direct scoring. The study provides an initial implementation and evaluation of Visual Jev as a reference-bound reward for multi-subject image generation.

0.21

Why VLMs Miss Small Objects, and When Zooming In Is Safe

cs.CV ·2026-10-07 ·Junzhe Shi, Yuan Gan, Shida Jiang
Abstract

Screenshots, documents, engineering drawings, shelves and maps hold many small objects of kinds that change from task to task, and vision-language models (VLMs) promise to find them from an instruction and an example. Yet they see each image through a fixed token budget and miss small objects, so every user, and every agent that calls a model, must choose whether to send an image whole, raise its resolution, split it, or zoom. We show that VLMs miss small objects for two reasons a caller can estimate before any call: an object gets too few visual tokens (S), and a call must search and report too much (L). Raising S by a factor a multiplies the tokens needed to see an image by a^2, while splitting at the same scale lowers L at a cost at most linear in the number of views. Under two monotonicity assumptions, splitting cannot lower expected recall if no view zooms out and views overlap by one object, and neither part can be dropped; every such split pays a minimum token cost, and a simple rule pays it up to an explicit factor. Better models need less S, but an image can be arbitrarily large while a call's budget is fixed, so splitting does not become obsolete, for zoom agents either. In about 187,000 requests to 12 VLMs, up to GPT-6.1, Claude Opus 5.5 and Gemini 4, on controlled images 24 of 40 predicted orderings hold significantly and none is violated; splits that never zoom out raised recall on average over models wherever S was below 1.5, and splits that zoom out lowered it in 19 of 21 model and data settings. The rule raised recall significantly in five of ten settings, by up to 0.28 (from 0.62 to 0.90), and precision and F1 in six, at a token cost known before any call.

0.21

Temporal Residual Bottleneck for Robust Asynchronous Collaborative Perception

cs.CV ·2026-10-07 ·Melih Yazgan, Ahmed Abouelazm, J. Marius Zöllner
published
Abstract · Accepted to ACCV 2026

Collaborative perception extends the sensing range of autonomous vehicles, but its performance degrades when shared features arrive stale or incomplete. Most latency-robust methods compensate delayed collaborator features through flow-guided alignment or direct feature transport. In this work, we formulate asynchronous collaborative perception as temporal residual prediction. Our Temporal Residual Bottleneck keeps a deterministic pose-warped collaborator feature as a conservative anchor and uses a $Δt$-conditioned xLSTM to extract residual temporal evidence from the available history. A detector-facing residual bottleneck then applies only gated, regularized corrections before ego-side fusion, reducing the risk of overwriting reliable static structure when temporal correspondence is uncertain. Experiments on DAIR-V2X and OPV2V show that our method is especially effective under severe fixed/irregular delays and packet drops. On DAIR-V2X, the reported checkpoint trades a small amount of synchronized peak accuracy for better robustness under stronger communication degradation. Controlled diagnostics further indicate that direct feature transport has oracle headroom but can become unreliable when deployed without accurate correspondence. These results support temporal residual fusion as a practical alternative for asynchronous and incomplete collaborative perception. Code will be publicly released at https://url.fzi.de/8dk38.

0.21

Visual Evidence Under Cross-Examination: Evaluating and Controlling Decision-Level Evidence Use in Vision-Language Models

cs.CV ·2026-10-07 ·Huiyao Zhang, Jin Bai, Zilong Su, Rui Guo, Chaofan Qin +4 more
benchmark
Abstract

Vision-language models increasingly reason through crops, regions, and tool-produced observations. Yet an observation can influence the answer without benefiting the candidate it supports. We study candidate-bound visual contribution: valid evidence should help, invalidating its supporting relation should remove its additional effect, and valid rebinding should redirect that effect to the newly supported candidate. We introduce CROSS-Bench, a benchmark of 28,000 decision problems, with matched invalidation and rebinding tests on a dedicated evaluation subset. Our RIVET interface preserves evidence identity and uncertainty, composes a candidate-conditioned response, and separately controls its strength. Shared-evidence experiments show that task accuracy and evidence ownership can diverge. Under matched capacity and training, RIVET increases normalized effect transfer from 0.512 to 0.651 where clean evidence has a positive effect. The advantage persists on common evaluation examples and across repeated decision-layer fits. With evidence predicted from raw inputs, RIVET improves CROSS-Bench accuracy by an average of 5.70 pp across four frozen backbones, relative to the same models without auxiliary evidence. These results separate the utility of visual evidence from the candidate-specific destination of its effect.

0.21

Clarify, Then Focus: Statement Normalization for Conversation Analytics at Scale

cs.CL cs.AIcs.LG ·2026-10-07 ·Mikhail L. Arbuzov, Karan Dave, Evgeniya Dontsova, Yaodong Hu, Vincent Lao +3 more
Abstract · 20 pages, 1 figure

Enterprise conversation analytics asks many questions of millions of interactions. Each question can require reconstructing what people mean and identifying which information matters, repeating costly interpretive work across the same transcripts. We propose a simple principle: clarify the text, then focus the reader. Statement normalization transforms dialogue into short, speaker-attributed statements with source references and semantic tags. The statements make meaning more explicit; the tags support selecting evidence for a particular question. Downstream models can use the full representation or a relevant subset, depending on what helps them make the decision. In an offer-suppression task on customer-service calls, normalization improves a supervised classifier without selection, while weaker prompted readers benefit from both normalization and selection. A small model can learn the normalization contract, while lightweight encoders handle tagging and downstream decisions. Sharing this preparation across questions supports an inference pipeline built entirely from small models, making analytics over millions of conversations substantially less expensive.

0.20

EM-SNN: Efficiently Modulated Spiking Neural Network for Remote Sensing Image Dehazing

cs.CV ·2026-10-07 ·Jie Shao, Jiaqi Ma, Wenwen Min, Beihang Song, Ning Chen +2 more
Abstract

Although spiking neural networks (SNNs) provide an energy-efficient alternative to artificial neural networks (ANNs), their application to remote sensing image dehazing remains limited. A key challenge arises from the coupling between haze-induced high-frequency attenuation and discrete spike thresholding. This interaction suppresses weak responses and fundamentally limits the recovery of edges, textures, and fine details in spiking dehazing models. To address this challenge, we propose the Efficiently Modulated Spiking Neural Network (EM-SNN), a dedicated spiking framework tailored to remote sensing image dehazing. EM-SNN integrates a statistics-driven Threshold-Modulated Leaky Integrate-and-Fire (TM-LIF) neuron to adaptively compensate for haze-induced contrast compression, together with a Spike Sobel Modulation (SSM) module that enhances structural cues and reduces depth-wise attenuation during spiking feature propagation. By jointly modulating activation scales and structural representations, EM-SNN improves dehazing performance while preserving the inherent event-driven sparsity of SNNs. Experiments on HRSD, RICE, RRSHID, and SateHaze1K demonstrate that EM-SNN achieves competitive dehazing performance while consuming only one quarter of the energy of the strong ANN baseline SFRDP-Net.

0.20

Pooling Representation Autoencoders for Efficient Diffusion

cs.CV ·2026-10-07 ·Ramón Calvo-González, Youssef Saied, François Fleuret
Abstract

Representation Autoencoders (RAEs) generate images from pre-trained visual fea- tures, but their dense token grids make generative modeling expensive. Motivated by local feature correlations, we introduce PoolDINO, a learned affine pooling operator that merges neighboring tokens. Training the pooling operator jointly with the RGB decoder preserves the standard two-stage RAE procedure without a separate feature auto-encoder. On ImageNet-256, 4x token compression retains comparable generation quality under internal guidance, while 16x compression trades some quality for greater efficiency. At a fixed budget of 100 sampling steps, latent-sampling throughput increases by 3.7x and 9.0x, respectively, relative to the unpooled baseline. Classification and dense prediction evaluations show that comparable guided generation quality can coexist with weaker performance on other tasks.

0.20

DeltaSplat: Iterative Gaussian Refinement for Pose-Free Feed-Forward 3D Gaussian Splatting

cs.CV ·2026-10-07 ·Chanung Park, Seunghyeon Song, Joo Chan Lee, Eunbyung Park, Jong Hwan Ko
Abstract · 11 pages, 6 figures

Pose-free feed-forward 3D Gaussian Splatting (3DGS) reconstructs a scene from sparse, unposed images in a single network pass, removing the need for camera calibration and per-scene optimization. However, camera estimation errors propagate into the predicted Gaussians and compound the geometric and photometric inaccuracies of single-pass prediction. To correct these errors, we introduce DeltaSplat, a lightweight Gaussian refinement module for pose-free feed-forward 3DGS. It iteratively renders the current Gaussians at the input context views and predicts per-Gaussian updates from the resulting residuals. A 2D residual alone, however, underdetermines the 3D correction. DeltaSplat therefore conditions each update on per-pixel Plücker rays and rendered depth as a soft geometric prior. A dual-branch convolutional mixer efficiently encodes these inputs, and per-attribute heads decode the fused features into position, opacity, and color updates. The module adds only ~2.2% parameters to the backbone and remains fully feed-forward at inference. On DL3DV, DeltaSplat reaches 26.64 dB PSNR in the pose-free setting, improving its state-of-the-art backbone by 1.75 dB and surpassing even baselines supplied with ground-truth cameras; consistent gains hold across 6-24 views and all camera regimes.

0.20

SAVU-BENCH: A Real-World Benchmark for Spatial Audio-Visual Understanding

cs.SD cs.AI ·2026-10-07 ·Yu Chen, Ruihang Liu, Yangguang Xu, Xinyue Jiang, Mohammed Bennamoun +3 more
benchmark
Abstract · Under Review

Spatial audio-visual understanding requires models to recognize not only what is present, but also where events occur and how they relate across modalities. Existing benchmarks often rely on simulated scenes, evaluate isolated spatial skills, and provide limited diagnostic insight into failure modes. We introduce SAVU-Bench, a real-world benchmark that systematically evaluates spatial audio-visual understanding across three capability levels and seven evaluation tasks. We further introduce SAVU-Diag, a scene-linked diagnostic set that decomposes reasoning questions into their prerequisite grounding and alignment sub-tasks. Evaluation of 12 representative models on SAVU-Bench reveals that while visual spatial grounding is relatively mature, spatial perception involving audio remains a primary bottleneck. SAVU-Diag further demonstrates that most reasoning errors co-occur with failures on these prerequisite tasks, though reasoning gaps persist even when prerequisites are correctly resolved. Motivated by these findings, we introduce SAVU-EA, a training-free evidence-augmented baseline that makes spatial cues more explicit. While SAVU-EA substantially improves spatial grounding and joint matching, high-level spatial reasoning remains challenging. Our findings highlight the urgent need for both robust spatial audio perception and deeper integration of cross-modal spatial relations.

0.20

Knowledge-Guided LLM for Intent-Driven Service Orchestration in Integrated Terrestrial and Non-Terrestrial 6G Networks

cs.CL ·2026-10-07 ·Wajdi Hammami, Irfan Azam, Soumaya Cherkaoui
Abstract

Natural-language service intents in integrated terrestrial and non-terrestrial (TN/NTN) 6G networks can be incomplete, ambiguous, or inconsistent, making unconstrained LLM generation challenging for reliable network-facing orchestration. This paper proposes a knowledge-guided LLM framework that combines grammar-constrained semantic parsing with a knowledge-graph (KG)-driven clarification loop to produce validated, orchestration-ready service profiles across four representative service classes. On a 200-sample intent dataset, the framework achieves 89% classification accuracy and 95% policy validity with an average of 0.66 clarification turns per request. Compared with a Pure-LLM baseline, the proposed framework improves policy validity and contradiction detection while requiring fewer clarification turns, demonstrating the value of explicit knowledge-guided reasoning for robust intent processing in integrated TN/NTN 6G orchestration.

0.20

Lossy Compressive Text Autoencoders

cs.CL ·2026-10-07 ·Vinko Sabolčec, Angelos Katharopoulos, David Grangier
Abstract

Our work explores learning a compressed latent representation of text, at the intersection of data compression and representation learning. We propose an autoencoder architecture that performs residual downscaling and upscaling of hidden representations along the time axis, with a residual low-dimension discrete bottleneck. We analyze our approach for different quantization methods, training objectives, and datasets. For different levels of compression, we evaluate the similarity between the original and reconstructed text both at the surface-level (BLEU) and at the semantic-level (LLM-based judge). Additionally, we evaluate our models on downstream question-answering and semantic text similarity benchmarks. Our approach results in compressed representations which are on par with lossless text compression algorithms at 2.24 bits per byte on web text data, while having good reconstruction and downstream task performance.

0.20

Boundary-Free Contextual Biasing: Depth-Adaptive Gating and Reading-Space Matching for Unsegmented Languages

eess.AS cs.CLcs.SD ·2026-10-07 ·Muhammad Huzaifah, Yu Pan, Zachary Yeo, Ningjie Bai, Guangzhao Yang
benchmark
Abstract

Contextual biasing supplies an ASR system with a list of expected words at inference time, but existing methods rely on word boundaries that Japanese and Chinese do not provide. We present a boundary-free biasing decoder for frozen public CTC models, built on a character-level Aho-Corasick automaton, with no training and no second pass. Two evidence-based mechanisms replace the boundary: a depth-adaptive gate that sets how hard to push from match depth, and reading-space matching for when the audio is right but the characters are wrong. On Aishell-1 NE's hard R1 subset we reach 66.5% recall, above the trained CLAS baseline (64%), transferring to WenetSpeech and to a second architecture without retuning. We release the first open Japanese contextual-biasing benchmark, where biasing lifts rare-word recall by 25 points at precision above 97%, and still by 19 and 22 points against 1,000-word lists.

0.19

Detecting Adversarial Images through Response Profiles of Vision-Language Models

cs.CV ·2026-10-07 ·Arash Vashagh, Roozbeh Razavi-Far
Abstract

Adversarial perturbations can alter the predictions of frozen vision-language models (VLMs) while leaving their confidence and image--text similarity patterns seemingly plausible. We investigate whether we can identify adversarial inputs based on the broader way an image interacts with a collection of general semantic prompts. Our detector summarizes these responses using category-level statistics, relationships among prompts, deviations from clean reference distributions, and stability under weak image transformations, producing a compact response profile that is classified by a lightweight model while the VLM remains fixed. We evaluate the approach on multiple public image datasets, several CLIP-style visual backbones, and a range of gradient-based, optimization-based, automated, and spatial attacks. The detector achieves strong discrimination in attack-specific settings and retains substantial performance when evaluated on attacks not seen during training. Under a controlled detector-specific protocol, the response-profile representation outperforms the evaluated embedding-geometry baselines. Additional analyses show that the feature groups provide complementary information and that the method remains effective under variations in the prompt configuration. We also examine inference cost and performance against detector-aware adaptive attacks. Overall, the results indicate that response patterns across semantic prompts provide a useful complementary signal for adversarial image detection in frozen VLMs.

0.19

GRC-Net: Global Representation Consistency Network for Unsupervised Multimodal Anomaly Detection

cs.CV ·2026-10-07 ·Seyoung Jeong, Jong Pil Yun, Sang Jun Lee
Abstract · 5 pages, 3 figures, Under Review

Automated quality inspection is essential for ensuring product reliability in manufacturing.While image-based methods effectively capture appearance-related defects, these methods are limited in detecting structural and geometric anomalies, motivating multimodal approaches incorporating 3D information. However, existing methods mainly rely on local patch-level representations, which often lead to unstable reconstruction errors even in normal regions. To address this limitation, we propose GRC-Net, which integrates a global-attention MLP to enforce global representation consistency across patch embeddings with a stable reconstruction module to improve reconstruction stability. The proposed method captures holistic contextual information through a global token and suppresses reconstruction noise by minimizing discrepancies between original and predicted embeddings. Experiments on MVTec 3D-AD and Eyecandies demonstrate that GRC-Net consistently outperforms existing methods at both image and pixel levels. Qualitative results further demonstrate reduced reconstruction errors in normal regions and more distinct reconstruction differences between normal and anomalous regions.

0.19

Towards Explaining Query Expansion Performance in Information Retrieval

cs.IR cs.CL ·2026-10-07 ·Sourav Saha, Aditya Dutta, Soumajit Pramanik, Mandar Mitra
Abstract

Query Expansion (QE) techniques have long been widely used in Information Retrieval (IR) to address the vocabulary mismatch problem. They remain relevant in modern retrieval systems, including those based on large language models (LLMs). However, no single QE method consistently outperforms others across all queries. This work seeks to explain the variation in QE performance through two complementary perspectives. The first is the concept of an Ideal Expanded Query (IEQ)--a hypothetical query that maximizes retrieval effectiveness with a downstream BM25 retrieval model. The second is a separability perspective, which quantifies how distinctly relevant and non-relevant documents are scored for a given expanded query using Cohen's (d). We develop a separability measure and practical formulations to approximate the IEQ and investigate how these factors relate to retrieval effectiveness. Extensive experiments on the TREC Robust collection, TREC DL 2019-2022 passage collections, and TREC DL 2019-2020 document collections reveal several interesting patterns. In particular, we find that expanded queries that are closer to the ideal expanded query tend to achieve higher retrieval effectiveness. We further show that the separability of relevant and non-relevant documents provides a complementary perspective for understanding QE performance.

0.19

A Deafening Silence: Catastrophic Forgetting Lives in the Output Embeddings of Tokens the Data Never Speaks

cs.CL cs.AI ·2026-10-07 ·Jonghyun Han, Younghoon Song, Jongyoul Park
Abstract

Continual pre-training and fine-tuning in Large Language Models (LLMs) inevitably induce catastrophic forgetting, typically mitigated by replay using often-inaccessible original data. In this data-free regime, we analyze where forgetting occurs and why. Systematic parameter freezing across five settings up to 1.4B reveals that forgetting concentrates selectively in the output embeddings of tokens rarely seen in the new corpus, whereas the same sqrt(v-hat) band of the body is inert and new learning resides elsewhere. This localization is governed by the vocabulary deficiency of the corpus rather than the training mode, allowing pre-retraining risk ranking from token counts alone within a fixed base model. Mechanistically, absent tokens receive persistent one-sided softmax gradients that Adam's second-moment (sqrt(v-hat)) normalization amplifies into full-sized updates. We therefore propose an intervention: raising Adam's epsilon exclusively for the output projection during training. Across eight settings spanning 160M to 12B parameters and four model families, this removes 39.4% to 67.9% of forgetting across all seven stable configurations without degrading target learning or requiring per-model tuning. The defense combines additively or better with replay (79.8% on Qwen/Korean) and rescues released-head LoRA from a 23-fold forgetting surge. Because post-hoc editing of the drifted rows recovers under 5% of forgetting, the intervention must operate during training. Our findings indicate that a single-line optimizer adjustment may serve as the primary defense against catastrophic forgetting where the corpus starves the vocabulary.

0.19

Video-Conditioned Generative Joint 2D-3D Hand Motion Recovery

cs.CV ·2026-10-07 ·Chen Xu, Yunqi Li, Binbin Huang, Brent Yi, Shenghua Gao +1 more
Abstract · 20 pages, 6 figures

Recovering faithful 3D hand motion from video remains challenging due to frequent occlusions and incomplete visual observations, which make frame-wise pose estimates unreliable and temporally inconsistent. To address this problem, we propose JoHan, a unified generative framework that recovers hand motion directly from video sequences without relying on intermediate per-frame pose predictions. Trained from scratch, our model jointly generates aligned 2D and 3D local hand pose sequences by learning their temporal dynamics and cross-representation correspondence. The generated 2D trajectories exploit direct spatial and temporal cues from the 2D images to guide the following generative 3D motion reconstruction, while the learned motion prior promotes temporal consistency. Their learned 2D-3D correspondence further enables recovery of the hand's global position and orientation relative to the camera. Extensive experiments on challenging benchmarks demonstrate significantly improved accuracy and speed in local hand-pose and camera-space reconstruction. Notably, our method captures much better hand-motion dynamics, producing significantly smoother motion than previous methods while maintaining high per-frame pose accuracy.

0.19

Beyond Token Revision: Investigating Mask-and-Replace Diffusion for Zero-Shot Text-to-Speech

eess.AS cs.SD ·2026-10-07 ·Hounsu Kim, Joonyong Park, Yuki Saito, Satoru Fukayama, Juhan Nam
Abstract · 5 pages, 1 figure, 3 tables. Under review

Unlike autoregressive models, discrete diffusion-based models for zero-shot text-to-speech generate speech tokens in parallel and can revisit earlier predictions. Mask-and-replace training extends mask-only training by randomly replacing some tokens, and its gains are commonly attributed to self-correction, the ability to revise previously generated tokens. However, exposure to randomly perturbed context during training may itself improve generation, raising the question of whether these gains require inference-time token revision. To investigate this question, we use DeMaR, which combines mask-and-replace training with confidence-ranked mask-only sampling while preserving the total training corruption probability. Trained from scratch on LibriTTS, DeMaR achieves lower word error rates (WER) than autoregressive and mask-only diffusion baselines using the same speech tokenizer. This advantage persists when each token remains unchanged after first being unmasked. Matched training conditions on two heterogeneous speech tokenizers show that both noisy-context augmentation and replacement supervision improve WER under this restriction. These findings demonstrate training-side benefits of replacement beyond enabling inference-time token revision.

0.19

RT-DETR-World: Transferring Rich LLM Semantics to Real-Time Open-Vocabulary Detection

cs.CV ·2026-10-07 ·Yupeng Zhang, Ziyi Zhao, Juntao Cheng, Sheng Wang, Ningnan Guo +2 more
Abstract

Open-vocabulary detection (OVD) recognizes categories unseen during training through textual category queries, yet achieving strong generalization with real-time efficiency remains challenging. Beyond vocabulary scaling, zero-shot generalization may benefit from reusable visual--semantic cues learned from seen data, including attributes, actions, states, and contextual relations. Existing real-time OVD methods primarily emphasize vocabulary coverage and efficient region/query--text matching; under strict efficiency constraints, compact detectors may struggle to absorb rich instance semantics and scene context. We propose RT-DETR-World, a compact DETR-style detector that transfers the rich semantics conveyed by descriptions during training while retaining lightweight query--text matching at inference. We construct GroundingCapv2 with three levels of supervision: category names for standard OVD, object descriptions conveying instance-level semantics, and image descriptions conveying object relations and scene context. These descriptions serve only as training-time semantic supervision. To help the compact detector absorb these semantics, we propose Dual-Path Description Alignment (DDA), combining a deployment-consistent MiniLM pathway with a training-only LLM teacher. MiniLM provides query--category supervision and object-description alignment, while offline teacher features supervise matched queries and global visual representations at the object and image levels, respectively. All teacher features are precomputed, and the teacher-side modules are removed after training. We further propose Relation-Aware Negative Relaxation (RNR), which uses teacher-derived semantic similarities to relax related negatives while preserving exact positives. Experiments demonstrate competitive zero-shot accuracy and a favorable accuracy--efficiency trade-off. The code will be released.

0.19

Mitigating Accent-Language Confusion in Self-Supervised Speech Representations for Language Identification

eess.AS cs.CLcs.SD ·2026-10-07 ·Minu Kim, Jihwan Lee, David R. Mortensen, Shrikanth Narayanan
Abstract · Submitted to ICASSP 2027

Spoken language identification (LID) aims to recognize the target language regardless of accent. In practice, however, LID models fine-tuned from self-supervised speech representations frequently confuse accents with languages, misclassifying non-native (L2) speech as the speaker's first language (L1). We show that non-native speech representations lie between native target-language and native L1 poles, causing systematic misclassification. To address this, we introduce a geometric projection that estimates an L1-bias direction solely from native speech and removes it before the frozen LID head. Across five MMS-LID models and non-native corpora, this projection substantially improves target language identification for L2-accented speech while preserving predictions for native speech. These results show that accent-induced L1 bias can be corrected directly within the representation space without L2 training data or model adaptation.

0.19

ECHO: Embodied Camera Observations of Human Object Carrying

cs.CV cs.RO ·2026-10-07 ·Xuefei Sun, Lorin Achey, Kali Hamilton, Alberto Speranzon, Gregory Grebe +2 more
benchmark
Abstract

Embodied and assistive agents must do more than recognize objects: they must reason about where an object belongs given the layout of an environment and the habits of the people who live in it. Progress on this problem has been limited, in part because no dedicated benchmark or dataset exists to define and evaluate it. Existing RGB-D scan datasets reconstruct static rooms without human activity, while human-object-interaction datasets capture motion without a navigable, fully reconstructed scene or a ground-truth notion of an object's natural destination. We introduce contextual object placement as a benchmark task: predicting an object's destination during an observed object-carrying episode. To support this task, we present Embodied Camera observations of Human Object carrying (ECHO), a large-scale synthetic dataset that pairs dense RGB-D scans of indoor scenes with recordings of an embodied human carrying everyday objects to context-appropriate destinations. ECHO is the first publicly available dataset to combine reconstructed scenes, human activity, natural language, and contextual-placement annotations. It comprises 3,805 human-annotated episodes across 159 floors of 115 HM3D scenes, involving 198 distinct objects. Each floor includes a complete RGB-D scan with human-annotated room labels and a surface list. Each episode provides synchronized RGB-D encounter clips; 6-DoF camera, human, and object trajectories; start and destination surfaces; an action caption; and a human-written context: a single sentence describing the inhabitant's routine that implies the destination without naming it. We evaluate contextual object placement using input-masked probes and an end-to-end baseline. Results show that no single input modality is sufficient, highlighting the need to jointly reason over scene structure, human activity, and contextual knowledge.

0.19

ExperienceIndex: Artifact-Grounded Memory

cs.CL cs.AIcs.IRcs.LG ·2026-10-07 ·Peter Baile Chen, Geoffrey X. Yu, Xinming Liu, Samuel Madden, Dan Roth +3 more
Abstract

Knowledge-intensive tasks require answering many questions by reasoning about a shared corpus of artifacts (e.g., court cases, or scientific literature). As humans interact with these corpora, they naturally accumulate experiential knowledge about artifacts, enabling them to quickly identify the complete set of relevant artifacts for each new task. However, existing AI agents lack appropriate memory solutions to build or reuse such artifact-grounded experience, leading to lower answer quality and higher online cost. Existing memory solutions extract and reuse information from prior task-solving traces, but they primarily focus on user preferences, factual attributes, or abstract reasoning patterns rather than persistent artifact-specific knowledge. We introduce ExperienceIndex, a novel experience layer for AI agents that captures and reuses knowledge about artifacts based on prior reasoning traces. ExperienceIndex stores two complementary forms of experience: (i) single-artifact experiences that summarize an artifact's contribution to prior tasks and (ii) artifact-pair experiences that encode structural relationships discovered during past reasoning. Integrated as lightweight middleware, ExperienceIndex uses an experience retrieval mechanism to guide agents toward the complete set of relevant artifacts for new tasks, improving both answer quality and efficiency. Across diverse corpora and agentic solutions with different search frameworks, ExperienceIndex delivers consistent gains, raising answer quality by up to 11.0 points and reducing online dollar cost by up to 50.5%. We further demonstrate two benefits: (i) cross-task generalization, where experiences accumulated from text-to-SQL tasks transfer to factoid QA tasks over the same artifact corpus, and (ii) teacher-student learning, where experiences from a stronger model enable a weaker model to reach comparable performance.

0.19

RELATE: An Evaluation Framework for measuring Relational Orientation of Large Language Models

cs.CL ·2026-10-07 ·Shivam Shukla, Jihye Kim, Shubham Gaur, Mahnaz Roshanaei, Magy Seif El-Nasr
Abstract

Large language models (LLMs) are increasingly used for emotional support, raising concern that sustained use may draw users away from their real-world relationships. Yet existing evaluations primarily focus on the safety, empathy, or helpfulness of responses, leaving under-examined a relational question: where does the model orient the user for continued support? To address this question, we introduce relational orientation, a property operationalized through two non-exclusive dimensions: inward-facing (IF) language, which positions the AI as the user's ongoing source of support, and outward-scaffolding (OS) language, which encourages real-world human connection. Grounded in psychological and sociological literature, we formalize a taxonomy of relational orientation and present RELATE, a persona-conditioned framework for measuring inward-facing and outward-scaffolding language at the sentence level in multi-turn dialogues. RELATE pairs 76 help-seeking situations adapted from naturally occurring questions with three simulated user styles, providing 228 evaluation stimuli. In our experiments, we evaluate seven LLMs using dialogues with six assistant turns each, yielding 1,596 dialogues and 69,194 assistant sentences. We assess these sentences using a primary rubric-based LLM judge and apply a secondary judge to a subset. Under automated evaluation, we find that the proportion of sentences labeled as IF is higher at the sixth assistant turn than at the first, while the proportion labeled as OS is substantially lower for hesitant, indirect simulated users than for explicit, reassurance-seeking users. RELATE provides a reproducible framework and a sentence-level signal for auditing and steering the relational orientation of supportive LLMs.

0.19

Mechanics of Long-Context Hybrid Models Part 1.1: From Hybrid Attention to Hybrid Position

cs.CL ·2026-10-07 ·Xiaoran Liu, Ziwei He, Xipeng Qiu
Abstract · 60 pages, 36 figures, 25 tables, under review

The architectural design of Large Language Models (LLMs) is shifting from traditional full-attention-only models to hybrid models, which combine different attention modules to improve long-context efficiency and performance in length extrapolation and context extension. To explain why hybrid models work and how to design them better, we propose Mechanics of Long-Context Hybrid Models. As Part 1.1 of this series, we begin with hybrids of full attention and either sliding-window attention (SWA) or gated variants of linear attention (LA), represented by GLA and GDN. We first observe a Seesaw Effect in Context Extension: LA hybrids benefit more from long-context continual pretraining, whereas SWA hybrids perform better under length extrapolation. We attribute this behavior to differences in the positional inductive biases induced by these attention mechanisms. We find that SWA hybrids suffer from a Short-Context Learning Trap, Short-Window Weariness, and Long-Window Laziness, and require extended windows to enhance performance in continual long-context pretraining. For LA hybrids, we summarize the Matthew Effect of Hybrid Position Extrapolation and propose Sliding-Window Linear Attention, achieving 16$\times$ training-free length extrapolation while maintaining 100\% accuracy on NIAH-SK1 in 64k context length.

0.19

CORAL-KV: Compact Cross-Layer Prediction and Residual Coding for Autoregressive Video KV Caches

cs.CV cs.LG ·2026-10-07 ·Albert Ge, Akash Haridas, Utkarsh Saxena, Ruijing Yang, Ian Colbert +3 more
Abstract

Autoregressive video diffusion models maintain a growing attention key-value (KV) cache, making memory a key bottleneck for long video generation. Recently, clustering-based quantization has emerged as an effective method to compress video KV caches by grouping similar key and value vectors then quantizing residual errors. However, when later frames attend to the compressed history, small reconstruction errors can compound as the video grows. Beyond the within-layer similarity exploited by clustering, we find strong dependencies across cache layers in video models. This motivates Coral-KV - Calibrated Offline, Residual-coded, and Adapted Locally - a KV-cache compression framework that uses cache states in each layer to predict keys and values in subsequent layers, then quantizes the remaining prediction residuals. We construct compact block-diagonal-plus-low-rank (block-DPLR) predictors: block-diagonal components model channel relationships within each attention head, while a global low-rank term models dependencies between heads. Coral-KV exposes a flexible fidelity-memory trade-off: it can reuse its predictors across videos or refit lightweight block-diagonal predictors to the current video for higher fidelity. Across extensive evaluations on both HY-WorldPlay and MAGI-1, our reusable configuration improves PSNR over QVG-4 by 0.32 dB and 4.95 dB, respectively, while using 48% and 37% less cache memory. Against QVG-Pro, our video-adapted configuration improves MAGI-1 PSNR by 3.77 dB while using 33.7% less cache memory. Coral-KV also maintains higher reconstruction fidelity over extended generation and reduces temporal discontinuities while retaining comparable generation latency.

0.18

Grammar Concept Annotation at Scale: Deployed Fine-Tuned Small Language Models Outperform Prompted Frontier Models

cs.CL cs.AI ·2026-10-07 ·Marjan Celikik, Ana Peleteiro Ramallo, Javier Morales
Abstract

Corrective feedback is among the best-evidenced drivers of second-language acquisition, yet corrections delivered during lessons rarely accumulate into an actionable view of grammar mastery. Prompted frontier models can provide such a view from learner--tutor lesson transcripts, but they are costly at scale. We close this gap by fine-tuning Qwen3.5 small language models (SLMs) on filtered and rebalanced teacher-generated supervision, then deploying an efficient 0.8B model in an end-to-end grammar mastery tracker for all English learners on our platform. Internalizing the annotation contract into adapter weights enables pairing the 0.8B model with a compact matched prompt rather than verbose instructions. On two human-curated benchmarks, both the deployed 0.8B model and a 4B reference comparator outperform prompted GPT-5.4 and GPT-5.6 Sol in precision and recall under nested matching criteria of increasing strictness: concept, evidence span, and correctness. The deployed 0.8B SLM reduces serving cost by approximately 16$\times$. A feature-level online experiment shows significant gains in learner engagement ($+15.8\%$) and key business metrics, including scheduled hours ($+2.1\%$) and GMV from new lessons ($+13.2\%$).

0.18

From Chunks to Functional Evidence: Function-Aware Retrieval for EDA Documentation QA

cs.IR cs.AI ·2026-10-07 ·Xiaotian Qiu, Kairui Liu, Shi Chenyi, Jinyuan Deng, Qi Sun +1 more
benchmark
Abstract · 10 pages, 2 figures, including appendices

Retrieval-Augmented Generation (RAG) is widely used to ground answers in documents. For complex technical documentation, however, the primary bottleneck is often not model reasoning but a mismatch between a query and the way knowledge is organized for retrieval. This mismatch is pronounced in Electronic Design Automation (EDA) documentation, where the information needed for an answer is scattered across heterogeneous yet tightly coupled artifacts. We therefore redesign the basic retrieval unit of RAG. Instead of operating on isolated chunks or binary relations, we collect typed artifacts into EDA functional units. Each unit is recorded as a hyperedge with links to its source chunks. We then train an encoder to align queries with functional units and combine unit retrieval with direct chunk retrieval. After mapping the selected units back to their sources, a unified reranker chooses the evidence given to the generator. On the newly constructed EDADocEval-QA dataset, our method improves ROUGE-L by 37.1% over Chunk RAG and 55.6% over the strongest graph baseline. On the public ORD-MMBench benchmark, it improves ROUGE-L by 30.0% over the strongest baseline. These results support function-aware evidence organization in the evaluated EDA documentation settings.

0.18

Your Prompt Should Do More: Effects of Retrieval Instructions in Embedding Models

cs.CL ·2026-10-07 ·Amanda Myntti, Jenna Kanerva, Veronika Laippala, Filip Ginter
Abstract

Prompted embedding models have recently received increasing attention, particularly for retrieval, where detailed retrieval instructions are provided as part of the retrieval prompt. Several new datasets and studies have examined this setting, showing that the current embedding models often struggle to follow such instructions reliably. In this paper, we study the mechanism of how instructions actually affect the representations of retrieval queries in asymmetric retrieval tasks. We show that models can fail to follow even simple task instructions when query-side distractors are included in the evaluation. We hypothesize that this behavior is driven by the training setup of current embedding models and their evaluation, and show that fine-tuning with added query-side distractors leads to substantial improvements, with minimal effect on other tasks.

0.18

SANet: Selective Attention Network for Infrared Small Target Detection

cs.CV ·2026-10-07 ·Yingmei Zhang, Wangtao Bao, Qin Xiao, Yong Yang, Weiguo Wan +3 more
Abstract

Infrared small target detection aims to accurately identify and locate dim targets in complex backgrounds and supports applications such as maritime surveillance and military search and rescue. However, the small size and weak contrast of infrared targets make it difficult to balance detection accuracy and false alarms. This paper proposes a selective attention network (SANet) for infrared small target detection. A dual-path semantic-aware module combines standard and pinwheel-shaped convolutions to preserve local spatial consistency and capture broader contextual information. Spatial and channel attention further refine the features and improve target-background discrimination. To address the limitations of static skip connections in U-Net, a selective attention fusion module adaptively integrates features across scales using spatially varying weights. It selectively enhances salient regions and improves discrimination between true targets and false alarms. Experiments on three public benchmarks, NUAA-SIRST, IRSTD-1K, and NUDT-SIRST, show that SANet achieves competitive performance in intersection over union (IoU), normalized IoU, detection probability, and false alarm rate. Its IoU exceeds that of the second-best method by 1.93, 4.32, and 2.21 percentage points, respectively. These results support the effectiveness of SANet in dim-target perception, discriminative feature representation, and background suppression.

0.18

One-Shot Adaptive Segmentation For Scientific Images

cs.CV ·2026-10-07 ·Tejaswi V. Panchagnula, Allison M. Davis, Fengqing Zhu
Abstract

Scientific image segmentation methods rely on extensive annotation and task-specific training, limiting adaptation across imaging modalities and experimental conditions. We present a training-free, one-shot framework that specializes vision foundation models using a single annotated reference image. The framework combines DINOv3 representations with background-adaptive feature orthogonalization to suppress artifact-related feature directions, after which cosine similarity localizes candidate regions for SAM segmentation. We evaluate the framework on red-blood-cell microscopy, structured-illumination pool boiling, and chest radiography. Relative to the strongest baseline, the proposed method improves mean IoU by 5.91% and 78.62% on the microscopy and pool-boiling datasets, respectively, while achieving comparable performance on chest radiographs. These results demonstrate that one-shot reference conditioning can adapt general-purpose vision models to specialized scientific segmentation tasks.

0.18

Concentration, Not Uncertainty: Why Targeted Synthetic Data Doesn't Help Camouflaged Object Detection

cs.CV ·2026-10-07 ·Akshat Dobhal, Sanjay Singh
Abstract · 29 pages, 3 figures, 12 tables

Camouflaged object detection requires pixel-accurate masks, but obtaining such annotations is slow and costly, making synthetic training images an attractive alternative. Under a fixed generation budget, however, it remains unclear which real-image regions to target for synthetic data generation. We study an uncertainty-guided generation strategy that clusters the unlabelled real images, identifies clusters on which the model is least certain, allocates synthetic generation toward those clusters, and iteratively retrains the model. Across 103 training runs, uncertainty-based targeting does not outperform random allocation. Five independent controls further show that this null result is not an artifact: targeted training sets are measurably different from random sets, but the difference is explained by concentrating the generation budget rather than by where uncertainty is concentrated, as every concentration rule we test reproduces the effect and, on boundary accuracy, so does aiming at the clusters the model was most certain about. Separately, we find substantial data contamination in CHAMELEON, with 50 of its 76 images duplicated from training data despite the standard overlap check reporting zero overlap. Together, these results show that, under a fixed synthetic-data budget, budget concentration, not uncertainty-based targeting, accounts for the observed training-set effects.

0.18

A Probabilistic Perspective on Wasserstein-Based Evidential Uncertainty for Out-of-Distribution Segmentation

cs.CV ·2026-10-07 ·Arnold Brosch, Abdelrahman Eldesokey, Michael Felsberg, Kira Maag
benchmark
Abstract · 20 pages, 6 images, 3 figures

Semantic segmentation networks operate on a fixed set of classes and therefore fail when out-of-distribution (OOD) objects appear during deployment, a critical limitation for safety-critical applications such as autonomous driving. Reliably identifying OOD objects requires well-calibrated epistemic uncertainty, yet common softmax-based confidence scores remain overconfident, while Bayesian alternatives such as Monte Carlo dropout or deep ensembles require costly repeated forward passes. Evidential Deep Learning (EDL) offers an efficient alternative by modeling class probabilities as a Dirichlet distribution learned from a single deterministic forward pass. Existing EDL formulations rely on Euclidean objectives that push predictions towards the simplex vertices, encouraging overconfidence rather than preserving uncertainty for unfamiliar inputs. We instead employ Wasserstein-based objectives, which respect the geometry of the probability simplex, and study the influence of the Wasserstein order on segmentation accuracy and OOD detection within a unified evidential framework. We evaluate this framework on a convolutional (DeepLabV3+) and a transformer-based (SegFormer) architecture on the SegmentMeIfYouCan benchmark, including LostAndFound, RoadObstacle21, RoadAnomaly21, and Fishyscapes. Our results show the optimal Wasserstein order is architecture-dependent: second-order objectives dominate on the convolutional backbone, third-order objectives on the transformer backbone, and our framework surpasses comparable baselines on most metrics, with a single deterministic forward pass.

0.18

Expert Coupling in MoE Pretraining: Reducing All-to-All Overhead with Correlated Placement and Token Shuffling

cs.CL cs.DC ·2026-10-07 ·Radha Gulhane, Quentin Anthony, Beren Millidge
Abstract

Mixture-of-Experts (MoE) layers replace the feed-forward block of a Transformer with E expert networks, and each token is routed to k of these experts. Under expert parallelism (EP) the experts are distributed across GPUs, and every MoE layer runs all-to-all collectives in the forward and backward passes to dispatch tokens to their experts and then combine the results. On a cluster with 8 AMD Instinct MI300X GPUs per node, these collectives can take 45% of the training step at EP32 with top-2 routing and 60% with top-6 routing. We find that early in pretraining routers have already learned to assign tokens to experts in correlated patterns, both within a layer and across layers. At top-2, 0.8% of the expert pairs in a layer are selected together by 42% of tokens, and the experts a token selects at one layer predict the experts it selects at the next layer. We use these correlations to keep more token--expert assignments on the token's own GPU, which reduces communication across GPUs and across nodes. Correlated expert placement puts experts that are often selected together on the same GPU. Combined with a dispatcher that sends each token to each GPU once, it removes up to 58% of dispatched rows. Token shuffling applies when sequence parallelism shards tokens across the EP group. It moves each token to the GPU predicted to hold its next-layer experts during the reduce-scatter that follows attention. On one node this raises the share of token--expert assignments served on the token's GPU from 12.5% to 59%. In Megatron-LM, across EP degrees from 8 to 64 with top-2 and top-6 routing, the two methods reduce all-to-all time by 1.16-2.63X and end-to-end step time by up to 1.41X. Neither method changes the models' underlying routing decisions or expert parameters.

0.17

Argos: Adapt Rich Geometric Priors for Generalizable Online Scene-Change-Detection

cs.CV ·2026-10-07 ·Ruihan Xu, Jiae Yoon, Kaichen Zhou, Ue-Hwan Kim, Luca Carlone
benchmark
Abstract · More details on the project website: https://www.multyxu.com/argos/

Robots operating in dynamic environments require reliable detection of how their surroundings change over time. Existing learning-based methods largely rely on pairwise 2D image features, which struggle under large viewpoint changes and occlusions, are sensitive to noise, and show limited generalization across domains, while explicit 3D approaches typically require costly offline optimization. We show that the implicit 3D knowledge of Geometric Foundation Models (GFMs) provides a strong basis for addressing these limitations. We introduce Argos, which adapts GFM features for joint scene change detection and 3D reconstruction. To address data scarcity and take a step toward a foundation model for scene change detection, we introduce a large-scale benchmark comprising two synthetic datasets and one real-world dataset, and train jointly across diverse datasets to improve cross-domain generalization. We further introduce Argos-SLAM, a real-time system designed for robotics, which performs online change detection and change-aware 4D mapping. Across benchmarks, our framework substantially outperforms existing baselines, with gains of up to 42.01% in change IoU and 27.91% in F1, while supporting scalable deployment in changing real-world environments.

0.17

Training with Missed Targets in Generative Recommendation: Separating Supervision from Probability Competition

cs.IR cs.LG ·2026-10-07 ·Xuesi Wang, Yangbin Shi, Xiaolin Zheng
Abstract · 12 pages, 4 figures, 8 tables

Generative recommenders return a limited candidate set and may omit observed targets before reranking. A training strategy appends these missed targets to reranker training lists, although inference still ranks only original candidates. This operation simultaneously changes retrieved-target weight, adds supervision over appended targets, and makes the two groups compete for probability. An append/no-append comparison therefore cannot explain changes in returned-item rankings. We construct three matched losses that hold retrieved-target weight fixed while introducing appended-target supervision and group competition separately. The intermediate loss trains within both groups but normalizes them separately, preventing training-only targets from competing with inference candidates. Experiments with a released OneRec model and locally trained Amazon generators show that this competition can harm returned-item ranking. In four prespecified Amazon Video Games comparisons, removing it improved full-target normalized discounted cumulative gain (FT-NDCG) by 7.8--22.2\%; 95\% intervals over users and three of four intervals over training runs excluded zero. A conservative development-set rule selected appended-target training for two of three generators in one held-out category and rejected it for all three in another, avoiding a 1.7\% loss. Candidate completion should therefore be evaluated for each generator rather than applied automatically.

0.17

LighTROcc: Lightweight 4D Occupancy Forecasting via Instance-Centric 3D Gaussians

cs.CV ·2026-10-07 ·Hwanhee Jung, SeungHyeon Kim, Inkyu Koo, Qixing Huang, Sang Ho Yoon +1 more
Abstract

Forecasting future 3D occupancy from surround-view cameras is essential for autonomous driving, yet existing approaches rely on dense voxel or bird's-eye-view representations whose cost grows rapidly with spatial resolution and prediction horizon. Because these representations do not explicitly maintain object identities, they also struggle to preserve instance consistency over time. We present LighTROcc, a lightweight instance-centric framework that represents movable objects with a compact set of learned queries and predicts present and future occupancy in a single forward pass. LighTROcc localizes each query through attention-guided forward lifting, combining image-space cross-attention, query-specific depth, and camera geometry to estimate its 3D center. Each instance is modeled as a mixture of anisotropic 3D Gaussians and propagated across future steps using predicted displacements, producing continuous, temporally consistent occupancy forecasts. Experiments on nuScenes and supplemented nuScenes-Occupancy show that LighTROcc outperforms the evaluated dense and instance-wise baselines in instance-level forecasting accuracy while maintaining strong voxel-level occupancy quality. Across different model configurations, LighTROcc achieves a favorable balance between forecasting accuracy and computational efficiency, demonstrating the potential of compact instance-centric modeling for camera-based 4D occupancy forecasting.

0.17

LIFT-SE: Linguistic Inference Followed by Flow Transformation for Generative Speech Enhancement

eess.AS ·2026-10-07 ·Haoyin Yan, Chengwei Liu, Zheng Xue, Xiaotao Liang, Jifa Cai +2 more
Abstract

Generative speech enhancement (SE) is prone to linguistic hallucination when semantic constraint is unreliable under severe noise and reverberation. Moreover, approaches that generate discrete codec tokens are bounded by the quantization error of the codec decoder, regardless of token-prediction accuracy. We propose LIFT-SE, a two-stage generative framework that decouples linguistic inference from acoustic synthesis within QRes-Codec, which exposes a quantized latent and its residual-completed continuous form. The first stage predicts clean codec tokens autoregressively, conditioned on frame-aligned features from a self-supervised front-end distilled toward clean speech. The second stage applies conditional flow matching to transport Gaussian noise to the continuous latent conditioned on the predicted tokens, and the frozen decoder reconstructs the enhanced waveform. Discrete generation provides naturalness, while continuous refinement restores signal fidelity. Experiments on the DNS1 and URGENT benchmarks show that LIFT-SE attains favorable linguistic consistency under reverberant conditions together with competitive perceptual quality, and systematic ablations verify the necessity of both stages. Code will be released in the future.

0.17

Spatial Latent Reasoning for Embodied Reference Understanding

cs.CV ·2026-10-07 ·Ling Li, Jianhui Zhong, Wei Liu, Zheng Jiang aand Yuxuan Liu, Jingyu Li +1 more
Abstract

Pointing-gesture visual grounding requires connecting hand geometry with the visual identity and extent of a referred object. A central challenge for continuous latent reasoning is how to organize these complementary cues into useful intermediate supervision. We propose Spatial Latent Reasoning (SLR), a framework that structures this supervision around an ordered sequence of geometric and visual states. A spatial ray state is supervised by fingertip position and pointing direction, followed by four states aligned with target-region features. To construct the visual targets, we introduce parity pooling, which applies polyphase grouping to average region tokens on four interleaved spatial supports. All states are generated recurrently during training and inference; auxiliary annotations are required only during training. On EgoPoint-Ground, the framework improves mIoU over same-backbone supervised fine-tuning by 2.8, 17.5, and 21.1 percentage points on Qwen3.5-4B, Qwen2.5-VL-7B, and Qwen3-VL-8B, respectively, with improvements on both hard subsets. On YouRefIt, it achieves 77.6% precision at IoU 0.5, a numerical margin of 5.2 percentage points over the reported state of the art under differing evaluation protocols. Ablations support joint geometric and visual supervision on the standard and similar-object sets, and favor parity over three alternative pooling operators on the standard set. These results support task-structured supervision for continuous pointing grounding. We will release the code and supporting materials.

0.17

Disentangling Linguistic and Paralinguistic Information with Routed Sparse Autoencoders

cs.CL eess.AS ·2026-10-07 ·Beimnet Bekele Guta, Xiaoyu Yang, Guangzhi Sun, Philip C. Woodland
Abstract · In submission

Self-supervised speech encoders contain linguistic and paralinguistic information in a shared, entangled representation space. We combine a TopK sparse autoencoder with route-specific supervision and cross-factor adversaries. Across frozen SPEAR and WavLM encoders, independent probes show factor-specific retention and suppression: linguistic information remains stronger in the linguistic route, while paralinguistic factors, including speaker identity, emotion, and prosody, are retained in the paralinguistic route and substantially reduced in the linguistic route. The route organisation learned on LibriSpeech persists on MSP-Podcast without representation-side retraining. Feature-space route interventions further transfer the swapped factor while largely preserving the information carried by the unchanged route. These results show consistent route-selective separation across encoders, corpora, independent probes, and representation-level interventions.

0.17

Beyond Masks and Trajectories: Flow-Guided Latent Action Injection for Stable Surgical Video Generation

cs.CV ·2026-10-07 ·Tsz-Yui Qin, Siyu Zhou, Chi-Keung Tang
Abstract

Surgical video generation holds substantial potential for surgical education, simulation, and data augmentation, yet generating surgical videos with realistic and clinically plausible motion remains challenging. Most existing methods rely on auxiliary conditions, such as masks, trajectories, depth, or reference videos, to achieve visually plausible synthesis. Yet, these auxiliary conditions typically require additional manual annotation or specialized acquisition, making it difficult to scale such methods beyond small, curated datasets. This motivates the need for a reference-free architecture capable of generating high-quality surgical video without requiring auxiliary visual conditions at inference time. We propose FLAIR, a Flow-guided LatentAction Injection framework for Reference-free surgical video generation. FLAIR learns action priors from optical flow of real surgical videos, dynamically predicts corresponding latent action representation from an input prompt, and injects it into a frozen base model to generate surgical videos with improved action consistency. We further construct SurgActionClip-30K, the first large-scale surgical vision dataset comprising action-centric segmented clips and structured caption labels, addressing the persistent lack of fine-grained, action-centric surgical datasets. Lastly, we introduce SurgMetrics, the first surgical domain-specific evaluation metrics for quantifying the quality of generated surgical videos, addressing the persistent absence of clinically grounded evaluation standards in this domain. Extensive experiments demonstrate that FLAIR enables generating high-quality surgical videos using text-only inference without auxiliary conditions, and validation in SurgMetrics demonstrates its strength in alignment with human perception compared to traditional metrics.

0.17

STRIKE: Learning Visual State Transitions for Physical World Modeling

cs.CV ·2026-10-07 ·Wenbin Teng, Tianshuo Xu, Depu Meng, Yuelei Li, Quentin Herau +3 more
benchmark
Abstract

Physical world modeling requires predicting how interactions change a scene, not merely generating coherent motion. We propose STRIKE, a framework that separates visual state transition learning from dense video generation. We construct event-aligned supervision by extracting observed states from training videos and pairing them with transition descriptions and temporal offsets. An image-based transition model learns to predict the next scene configuration from the current image, a local transition specification, and elapsed time. At inference, a pretrained vision-language planner predicts time transition specifications, and recursive application of the learned transition model produces a sequence of future visual states. A separately trained dynamic model then generates the complete rollout conditioned on these states and their temporal locations. Experiments on Physics-IQ Verified, PhyGenBench, Pisa-Experiments, and RoboTwin2.0 show improvements of STRIKE over the corresponding video-backbone baselines in benchmark measures of physical consistency and manipulation-video fidelity. These results support learned visual state transitions as an effective intermediate representation for physical world modeling.

0.16

On the Necessity of Attention-FFN Split in Vision Transformers

cs.CV ·2026-10-07 ·Junhyeok Kim, Jinyeong Kim, Jae Wan Park, Seong Jae Hwang
Abstract

The standard Transformer architecture relies on a rigid pattern that alternates Attention and Feed-Forward Network (FFN) layers. Despite its widespread adoption, the inductive bias imposed by this strict separation has not been systematically examined. In this work, we investigate the necessity of the Attention-FFN dichotomy in Vision Transformers (ViTs). To facilitate this analysis, we introduce the AttenFeed module, a unified component that integrates the functional properties of both Attention and FFN. Based on this module, we devise the unified Vision Transformer (uViT), which replaces the conventional alternating Attention-FFN structure with a sequence of AttenFeed modules. We then use uViT as a control group that relaxes the Attention-FFN dichotomy of the standard ViT and systematically compare the two models across multiple datasets and model scales. Our experiments reveal that the Attention-FFN dichotomy can hinder performance at smaller model scales due to the rigid parameter allocation of ViTs. The AttenFeed module and uViT serve as new analytical tools for understanding the Attention-FFN structure and offer theoretical insights into the heuristically designed architecture of conventional ViTs.

0.16

Linear-Time Context Compression with State-Space Encoding

cs.CL cs.LG ·2026-10-07 ·Marc Serra Ortega, Andrey Barsky, Carlos Boned Riera, Tomás Ockier, Dimosthenis Karatzas
Abstract · 15 pages, 7 figures

Transformer-based large language models (LLMs) impose substantial computation and memory costs as context sizes increase, motivating methods that compress a context into a shorter representation before inference. Existing soft-prompt compression methods typically use the same Transformer backbone as used in the target model, where producing the compressed representation leads to significant latency overheads. This restricts them to the offline setting where context is compressed once and then reused across many queries. We propose the Boa Compressor, a plug-and-play online context compressor built on a Mamba state-space model (SSM) that encodes arbitrary-length context in a single linear-time pass, producing a representation on which we condition a frozen target LLM. The SSM is trained by distillation, matching the next-token logit distribution from the LLM ingesting the full context (as teacher) to the same LLM ingesting the compressed representation (as the student). At equal compression ratios, Boa matches or surpasses the performance of existing state-of-the-art offline compression baselines, while being an order of magnitude faster than Transformer-based implementations, thereby reducing end-to-end latency rather than adding compression time overhead. We demonstrate that Boa also scales well to longer contexts and to compression ratios as aggressive as 16x. By enabling effective compression that is cheap enough to run on every query at inference time, Boa becomes a practical and general tool for efficient long-context inference.

0.16

Learning Multi-Step Query Rewriting via Corpus Feedback for Conversational Search

cs.IR ·2026-10-07 ·João Coelho, Hong Wang, Jie Yuan, Zhuoer Wang, Samson Koelle +1 more
Abstract

Conversational Query Rewriting (CQR) turns a context dependent user turn into a standalone query for a retriever, and most methods do this in a single step from the dialogue history before retrieving once. The rewrite is therefore fixed before any corpus evidence is available to correct its reference resolution or its vocabulary. We recast CQR as a sequential retrieval problem: an agent rewrites the current turn, retrieves, and conditions its next rewrite on the returned passages. The agent acts in a typed space of three rewriting operations, resolving conversational intent into a standalone query, generating lexical reformulations, or synthesizing pseudo-documents for document-to-document matching, together with a stop action that ends the episode. We train the policy with supervised fine-tuning followed by reinforcement learning against a single retrieval-quality reward, using no human rewrite annotations. Across TopiOCQA and QReCC, the agent outperforms several retrieval-aligned baselines, while remaining effective across retrieval backends and generalizing to the CAsT benchmarks without additional training. Further analysis shows that, through retrieval-reward optimization alone, the learned policy develops a behavior of grounding pseudo-documents in passages retrieved by earlier steps, substantially improving retrieval.

0.16

OnlineQAT: On-Policy Distillation for Ultra-Low-Bit Large Language Models

cs.CL cs.AIcs.LG ·2026-10-07 ·Wenjun Wang, Heng Li, Yanggan Gu, Hongxia Yang
Abstract

Quantization-aware training (QAT) can recover much of the accuracy lost when large language models are compressed below four bits. Existing re- covery stages, however, are commonly optimized on fixed completions or teacher-generated answers, whereas the deployed quantized model condi- tions on prefixes generated by itself. Quantization errors can therefore move the model into states that are absent from offline recovery data. We introduce OnlineQAT, a two-stage framework that first obtains a usable low-bit initialization through block-wise QAT and then performs on-policy distillation (OPD) on student-generated responses. At each visited pre- fix, a frozen full-precision teacher provides a sampled reverse-KL training signal. On Qwen3-1.7B, OnlineQAT obtains the best average among the compared quantized methods: 57.28 at W3A16 and 32.52 at W2A16, im- proving over ReasoningQAT by 2.90 and 0.44 points, respectively. The results suggest that student-visited states provide a useful recovery signal beyond fixed-completion training, particularly at three bits.

0.16

trACT: temporal revelation Airborne Camera Trap

cs.CV ·2026-10-07 ·Oliver Bimber, Rakesh John Amala Arokia Nathan, Mohamed Youssef, Vinayak Lal Bhatnagar, Ralf Berger +1 more
Abstract

Effective remote monitoring and surveillance using drones are frequently impeded by severe environmental and thermal clutter, dynamic vegetation, target camouflage, and system latency. Drawing inspiration from the hunting strategies of birds of prey that hover and stabilize their vision to isolate subtle ground motion, we introduce trACT (temporal revelation Airborne Camera Trap), a lightweight, real-time aerial robotics framework designed for autonomous consumer drones. The system integrates Temporal Max Pooling (TMP), a low-level signal processing method that transforms imperceptible movement across a rolling integration window into robust value and time encodings, with self-supervised motion anomaly detection to isolate target motion from background environmental motion caused by wind gusts and drone drift. To overcome mechanical and processing delays, trACT combines motion prediction with automated gimbal-stabilized optical zoom verification and equitable multi-target verification balancing. Extensive real-world field experiments in densely forested wildlife habitats and surveillance scenarios demonstrate that trACT successfully bridges the gap between wide-area aerial monitoring and precise, autonomous target verification under challenging operational conditions.

0.16

Sparse Attention Is Matrix Approximation, Not Choosing from a Bag of Values

cs.CL ·2026-10-07 ·Fang Wan, Xufeng Liu, Fan Li, Yi Liu
position
Abstract

Large Language Models (LLMs) achieve strong performance across many domains, but their efficiency is limited by the quadratic cost of attention with respect to prompt length. Sparse attention reduces this cost by retaining only a small fraction of query-key interactions to approximate the full attention matrix. However, existing methods are trapped in a mathematically wrong view: they simply keep large scalar entries or high-mass regions of the attention matrix. This treats the attention matrix as a bag of values, ignoring that it is used as a structured matrix whose entries jointly determine the attention output through multiplication with value vectors. We argue that this is the core conceptual issue: sparse attention should be formulated as matrix approximation, not as blindly choosing the largest values from a bag of entries. Based on this view, we propose Matrix Approximation Sparse Attention (MASA). MASA replaces raw attention-mass ranking with a closed-form score that measures how much each sparse unit reduces matrix-product approximation error. As a theory-grounded plug-in correction, MASA can be added to existing sparse attention frameworks without changing their sparse kernels or budgets. Extensive experiments across multiple sparse attention methods, benchmarks, and LLM backbones show consistent accuracy gains, supporting both MASA and the matrix-approximation view of sparse attention.

0.16

FedSSMCoOp: SSM Encoders for light-weight Federated Prompt Learning for Few-shot Classification

cs.CV ·2026-10-07 ·Ankita Das, Ambarish Parthasarathy, Sumohana S. Channappayya, C. Krishna Mohan
Abstract

Vision-Language Models (VLMs) have shown strong performance across a wide range of downstream vision tasks, thanks to the complementary information contained in the respective domains. Despite the performance gains, most of these approaches rely on aligning these domains using the cosine similarity metric, which fails to capture token-level structure and cross-modal interactions prior to the classification stage. This is especially critical in biomedical applications under federated constraints, where data sharing is restricted, labeled data is scarce at each site, and it differs widely across institutions, leading to substantial statistical heterogeneity. To overcome this issue, we propose FedSSMCoOp, a federated few-shot image classification framework that enables multimodal learning while preserving data privacy. With the help of the SSM-based Vision Mamba and Cross Mamba blocks, and by optimizing only the soft-prompt and communication-prompt updates in the federated setting, the framework prioritizes both computation and performance. Importantly, this eliminates the need to use an external Large Language Model (LLM) for feature alignment. The framework is further trained and evaluated on various biomedical image datasets, and its performance is assessed. The proposed framework delivers stable performance relative to the baselines and is, on average, 1.96 times lighter. The corresponding script will be made available soon.

0.16

Stochastic Teacher Intervention for Agentic On-Policy Distillation

cs.CL cs.AI ·2026-10-07 ·Junnan Liu, Linhao Luo, Zhijun Chen, Qianren Mao, Thuy-Trang Vu +1 more
benchmark
Abstract · Work in progress

On-policy distillation (OPD) efficiently transfers capabilities from a stronger teacher to a student language model through dense token-level supervision on student-generated rollouts and has shown promise on complex tasks such as mathematical reasoning. However, in multi-turn agentic tasks, student decisions shape subsequent observations, causing early errors to accumulate across turns. The resulting trajectories can drift away from the teacher's rollout distribution, making the teacher's token-level supervision less reliable or even counterproductive for OPD training. To address this issue, we introduce STI-OPD, a stochastic teacher intervention framework for multi-turn agentic OPD. During multi-turn interaction, STI-OPD uses teacher intervention guided by teacher-student policy discrepancy to replace the student's proposed action with a teacher-generated one to maximize the acquisition of reliable supervision. We further develop a stochastic intervention strategy, addressing the limitations of previous threshold-based or fixed-schedule approaches, that estimates policy discrepancy using KL divergence and maps it to an intervention probability. By sampling whether to intervene from this probability, STI-OPD adaptively balances teacher control with student exploration. To learn from the resulting mixed-policy trajectories, we introduce an Importance-Weighted Reverse KL objective that corrects the token sampling mismatch between teacher-generated responses and the student policy to preserve the original OPD objective. Across tool-integrated reasoning and long-horizon interaction, STI-OPD outperforms the strongest prior OPD baseline on every evaluated benchmark and student size. Ablations further show that both discrepancy-guided intervention and importance weighting contribute to these gains.

0.15

Perceptually Aligned Evaluation of Style Transfer

cs.CV ·2026-10-07 ·Yang Deng, Eleftherios Ioannou, David Mould, Steve Maddock, Paul L. Rosin +1 more
benchmark
Abstract

Style transfer lacks a reliable evaluation standard: ground truth is inherently ill-defined, and existing automatic metrics often fail to reflect human preference. This paper introduces ASTRA (Assessment of Style TRansfer Algorithms), an approach for automatic evaluation of style transfer algorithms; it contains two components, ASTRA-Data and ASTRA-Score. ASTRA-Data consists of a benchmark image set of content and style references, a collection of style transfer results generated on the benchmark set, and user study data capturing human judgements through a two-stage pairwise comparison protocol. From these annotations, we derive ranking-based ground truth for content preservation, style fidelity, and overall preference. Based on ASTRA-Data, we construct ASTRA-Score, a learnt evaluator that predicts preference-aligned scores from content-style-stylization image triplets, enabling automatic and scalable evaluation of new models applied to the benchmark set. Experimental results demonstrate that ASTRA-Score achieves substantially higher correlation with human rankings compared to prior metrics. Overall, ASTRA establishes a robust mechanism for standardised evaluation of style transfer methods.

0.15

Judging in Latent Space: Efficient Generative Reward Modeling via Semantics-Preserving Compression

cs.CL ·2026-10-07 ·Mingqing Yuan, Xiaobo Liang, Junwei Yang, Ziwei Chen, Zeren Zhang +3 more
benchmark
Abstract

Reward modeling often requires jointly representing and reasoning over multiple evaluation criteria, yet verbalizing this process token by token can incur substantial inference cost. Recent work on latent reasoning suggests that continuous states may support this computation more compactly. We introduce LatentGRM, a latent evaluation framework built on semantic chunking, compression, and reconstruction. By using the structure of rubric-guided evaluations to guide compression, LatentGRM learns compact continuous trajectories that support autonomous pairwise judgments without generating textual assessments. A separate interpreter reconstructs evaluation text from these trajectories, providing an offline view of the information retained under compression. Under matched training data and backbones, LatentGRM achieves competitive aggregate preference accuracy relative to explicit Supervised Fine-Tuning (SFT) judges at both 4B and 8B scales. Across four benchmark domains, LatentGRM-8B compresses evaluation trajectories by 8.9--9.2x and reduces total judge inference time by 6.1--7.0x at vote@5. Controlled rubric interventions show that criterion-dependent preference information is carried through the latent sequence. Together, these results demonstrate that continuous latent evaluation can substantially reduce inference cost while preserving competitive judgment quality.

0.15

Training-Free Instruction TTS Gender Bias Calibration Using Model-Adaptive Steering

eess.AS cs.SD ·2026-10-07 ·Kuan-Yu Chen, Yi-Cheng Lin, Jeng-Lin Li, Jian-Jiun Ding
benchmark
Abstract · 5 pages, 1 figure, 3 tables, submitted to ICASSP 2027

Instruction text-to-speech (ITTS) systems systematically encode acoustic gender skews from descriptive style prompts, such as occupations or personas, even when demographic attributes are left unspecified. Calibrating the implicit gender distribution of the synthesized voices, across heterogeneous architectures and without model retraining or overriding explicit user prompts, is an open problem. In this work, we propose \textit{model-adaptive steering}, a training-free bias calibration method that steers post-encoder conditioning representations using a group deviation vector paired with a coarse-to-fine strength search on a development set. A deterministic lexical gate bypasses intervention whenever explicit gender keywords are detected, preserving intended prompt semantics. Evaluated on a 12,800-prompt held-out benchmark across four ITTS models (three architectures), the proposed method reduces aggregate calibration error from 11.5--28.1 to 0.8--5.9 percentage points (e.g., shifting Parler-TTS Mini from 78.1\% and PromptTTS++ from 27.3\% female to 50.8--52.4\%). Output rates stay within 0.9 pp across anchor set sizes at fixed operating points, while UTMOS decreases by at most 0.08 and WER degrades by at most 3.3 pp.

0.15

Real Long-Term Memory for AI: A 50-Million-Token Window That Is Faster and Cheaper Than Recompute

cs.CL cs.AIcs.DCcs.LG ·2026-10-07 ·Sietse Schelpe
Abstract

A large language model can only use the text that fits in its context window, and it recomputes its internal key-value (KV) state for a prompt every time the prompt is sent. We test a memory layer, the public package galahad-kv, that saves the KV state of each block of about 16,000 tokens to encrypted local NVMe disk and loads it back later, byte-exact, without recomputing it. We ran it on 50,000,000 tokens of real public text, served through vLLM on one NVIDIA H100, with Gemma 4 12B and Gemma 4 31B. Every block we probed was loaded back from the encrypted store with no recompute (100 of 100, at depths from 0 to 50M tokens) on both models. Loading a block was 2.8x to 4.3x faster than recomputing it and used 8.8x to 12.3x less GPU energy, and GPU memory stayed flat over the whole 50M-token stream. Asked about facts planted millions of tokens earlier, the 12B model gave the right answer 82 times out of 100 and the 31B model 98 times out of 100. Neither model made up an answer. The limits are as follows. This is reuse of stored state, not a wider attention window: one block is loaded at a time, and how well a question is answered depends on the model. Writing the memory is a one-time cost, and the store takes terabytes of local NVMe disk. We describe the test protocol, which is built to resist common ways of gaming long-context benchmarks, and give a single-GPU reproduction that uses public software and a free licence for the package.

0.15

AdSpark: A Large-Scale Dataset and Benchmark for Product-Centric Advertisement Video Generation

cs.CV ·2026-10-07 ·Zhifei Yang, Zhao Jiang, Keyang Lu, Honghe Zhu, Zheng Zhang +4 more
benchmarkdataset
Abstract

Product-centric advertisement video generation aims to create promotional videos that preserve fine-grained product identity while presenting selling points through coherent multi-shot narratives. However, this emerging task remains underexplored due to the lack of large-scale advertisement-specific datasets and comprehensive evaluation frameworks. To address this gap, we introduce \textbf{AdSpark}, a large-scale dataset and benchmark for product-centric advertisement video generation, based on data from a major e-commerce platform. \textit{AdSpark-300K} contains approximately 300K reference image--prompt--video triplets, comprising a real-world subset and a synthetic subset. Each sample provides structured advertisement annotations, including product identity annotations, selling-point descriptions, creative plans, and aligned audio scripts, enabling models to learn product preservation and advertisement-oriented visual storytelling. We further propose \textit{AdSpark-Bench}, a diagnostic benchmark that evaluates generated advertisements across six dimensions, including visual quality, product fidelity, instruction adherence, temporal coherence, audio alignment, and advertisement effectiveness. Based on AdSpark-Bench, we evaluate representative models, revealing key challenges in product preservation, multi-shot storytelling, and selling-point visualization. Experiments with AdSpark-300K-finetuned models further validate the effectiveness of our dataset. AdSpark provides a unified dataset and benchmark for future research, and we will release the dataset upon acceptance.

0.15

Adaptive Visual Token Reduction for Accelerated Image Understanding

cs.CV ·2026-10-07 ·Seyoung Jeong, Jong Pil Yun, Sang Jun Lee
Abstract · 5 pages, 2 figures. Under review

Large Vision-Language Models achieve strong VQA performance, but processing high-resolution, information-rich images requires substantial computation, motivating visual token reduction. However, existing methods often prune individual tokens or rely on fixed-size cropping, limiting their ability to preserve spatially structured information such as horizontally or vertically elongated text. To address this limitation, we propose ReFIT, an instruction-guided visual token reduction framework for efficient LVLM inference. ReFIT consists of Relevance-Guided Window Reshaping (RWR) and Instruction-Guided Token Refinement (ITR), where RWR captures instruction-relevant regions by adapting to their spatial characteristics, while ITR further removes unnecessary visual tokens. Experiments on four VQA benchmarks demonstrate that ReFIT improves answer accuracy while reducing computational cost, and qualitative results demonstrate its effectiveness in localizing relevant regions and removing unnecessary visual information.

0.15

Quantifying Volumetric Risk: Class-Aware Asymmetric Weighted Conformal Prediction for 3D Medical Image Segmentation

cs.CV ·2026-10-07 ·Shadi Alijani, Fereshteh Aghaee Meibodi, Homayoun Najjaran
benchmarktheory
Abstract

Reliable volumetric segmentation is critical for clinical diagnostics, yet foundation models such as MedSAM remain deterministic and lack calibrated uncertainty under distribution shift. Existing conformal prediction methods offer statistical guarantees but are frequently applied in 2D and assume symmetric error distributions, so they do not capture the class-specific biases that arise in 3D multi-class segmentation. We propose Class-Aware Asymmetric Weighted Conformal Prediction (CA-WCP), which combines latent-space density-ratio weighting for covariate shift with directional quantiles for the lower and upper volume bounds, and scales each bound by a class-specific asymmetry factor derived from validation-set false-positive and false-negative rates. We prove that CA-WCP retains the weighted-exchangeability marginal coverage guarantee for every class, and we evaluate it on 3D brain tumor segmentation (BraTS 2020) and on a synthetic multi-organ CT benchmark constructed under covariate shift. On both benchmarks the 95\% Clopper--Pearson interval for the observed coverage of CA-WCP contains the nominal 90\% level for every semantic class, while interval width is reduced by 8--14\% relative to symmetric weighted conformal prediction. We further encode the calibrated intervals into structured prompts for a multimodal large language model to produce uncertainty-conditioned radiology reports, linking distribution-shift-aware uncertainty quantification to interpretable clinical communication.

0.15

One Frame, Full Heartbeat: ECG-Free Cardiac Cine MRI Synthesis via Phase-Conditioned Flow Matching

cs.CV ·2026-10-07 ·Shiyi Wang, Ruochen Sun, Peirong Liu, Xiang Li, Fangxu Xing
benchmark
Abstract

Cine cardiovascular magnetic resonance (CMR) analysis relies on multi-frame sequences capturing the full cardiac cycle. However, standard multi-frame acquisition depends heavily on electrocardiogram (ECG) gating and repeated breath-holds, posing challenges in uncooperative populations, resource-limited settings, and temporally corrupted datasets. Existing methods that synthesize full cardiac sequences either rely on explicit ECG signals to parameterize myocardium function, or employ deformable registration without physiological constraints, failing to faithfully reproduce clinically relevant dynamic metrics such as ejection fraction (EF) and ventricular contraction magnitude. We present PhaseFlow, a unified generative framework that overcomes both limitations. PhaseFlow estimates a non-linear cardiac phase signal directly from the input sequence via a segmentation-derived left-ventricular (LV) area curve, capturing the asymmetric dynamics of systole and diastole without any ECG dependency. At inference, this phase signal is provided by a pathology-specific template, informing phase-specific frame generation. A rectified flow model conditioned on the phase and slice position synthesizes the full cardiac motion trajectory in the latent space, decoded into a diffeomorphic displacement field that warps end-diastole pixel intensities directly, eliminating the reconstruction blur often accompanying the variational autoencoder. On the ACDC benchmark, PhaseFlow achieves superior physiological fidelity and image realism, with best LV volume curve $R^2$, structural similarity (SSIM) and generative quality (FID) among all baselines. Ablation studies confirm that each proposed component contributes measurably to the overall performance.

0.15

It Is Not Seeing the Hazard: A Frozen Vision-Language Safety Score Measures Its Caption Bank

cs.CV cs.LG ·2026-10-07 ·Samuel Tetteh, Cody Fleming
Abstract

Frozen vision-language models increasingly provide safety signals for reinforcement learning. Their use assumes that similarity to language describing danger indicates the hazard itself. Yet policy return and collision rate cannot reveal whether a score detects hazards or responds to correlated features of the scene. VLM-based methods have reported gains in driving and safe-RL benchmarks by converting image-text similarity into rewards, costs, or confidence weights. Such signals promise to reduce reliance on manually designed feedback. They may also reflect prompt structure, embedding geometry, or camera viewpoint, leaving their safety meaning unverified. To address this gap, we present a controlled evaluation of a frozen CLIP prompt-margin safety score. We apply the score to trajectories generated by policies that never receive it, match pre-contact observations to contact-free observations with comparable hazard geometry, and vary the captions, encoder, and camera view. Across three policies, 180 episodes, and 130 isolated contact onsets, the score decreases for about twenty steps before contact. Mechanism controls indicate that the score mainly tracks resemblance to the scene shared by its captions and changes with caption separation and camera view. A constant-confidence control retains the lower catastrophe-rate point estimate, so policy gains do not establish hazard perception.

0.15

TerraAgent: Demonstration-Grounded Multi-Agent Reasoning for Long-Context Remote Sensing Visual Question Answering

cs.CV ·2026-10-07 ·Jingwen Huang, Yiming Fang, Hanxiao Wu, Boyuan He, Keyi Luo
Abstract

Remote sensing visual question answering (RS-VQA) requires answers that truly depend on image evidence, yet current remote sensing multimodal large language models (RS MLLMs) are still evaluated mainly by accuracy, which cannot reveal whether a correct answer was produced without using the image. This setting introduces three challenges: demonstrations retrieved across different ground sample distances (GSD), sensors, and city textures may be poorly matched; real queries often involve long visual contexts from multi-scale tiles, multi-temporal frames, and retrieved knowledge, where evidence can be diluted and errors can propagate through multi-step reasoning; and prior-only answering remains invisible to standard metrics. We present AVA, a training-free framework that wraps a frozen RS MLLM with three modules. GDR performs GSD-bucketed, scene-graph-based demonstration retrieval and filters demonstrations whose answers are unsupported by their images. MACE decomposes each query into an evidence-typed sub-question DAG, routes sub-questions to suitable perception tools, and records verified provenance in a shared evidence ledger with bounded re-perception. CVA audits the final answer through counterfactual perturbations of cited evidence, including ablation and swap, and computes a visual-dependency score for confidence calibration and evidence-forcing re-query. On RSVQA-LR/HR, VRSBench, and LHRS-Bench, AVA improves average accuracy by up to 2.40 points over the strongest same-setting baselines, raises the visual-dependency score from 0.28 to 0.49, reduces the prior-only answer rate from 46.2 percent to 21.7 percent, and lowers the hallucination rate from 17.1 percent to 9.8 percent, with 2.6x the cost of a single backbone forward pass and no fine-tuning.

0.15

TraceICL: Thread-Resolved Demonstration Curation with Consistency Calibration for Robust Multimodal In-Context Learning

cs.CV ·2026-10-07 ·Yigit Yilmaz, Elena Petrova, Mehmet Kaya, Lucia Rossi, Amir Rahman
Abstract

Interleaved multimodal in-context learning (ICL) lets frozen large vision-language models (LVLMs) acquire new tasks from a handful of image-text-answer demonstrations, yet it remains fragile: demonstrations bind to the wrong images, redundant exemplars crowd out informative ones, and noisy answers are copied verbatim into predictions. We present AVA, a training-free framework that organizes the demonstration pipeline of multimodal ICL into three composable stages. A Thread-Resolved Decomposer (TRD) segments the interleaved context into per-sample threads and re-stitches them with explicit boundaries, restoring image-text binding. An Influence-scored Demonstration Selector (IDS) estimates each candidate's marginal influence with one extra forward pass, removes redundancy, and orders the survivors by descending value. A Weak-to-Strong Consistency Calibration (WSC) module scores each demonstration by a weak sibling model's reconstruction confidence and softly reweights its contribution during decoding. On five vision-language benchmarks with interleaved 4-shot prompts, AVA improves the strongest same-setting baseline by +1.7 points on average, with gains of up to +2.4 points on VizWiz, shrinks the require-visual performance gap on ScienceQA-IMG, and retains most of its gain under 30% noisy demonstrations, at a 26% latency overhead that an offline cache amortizes. No gradients are ever computed; the LVLM stays frozen.

0.14

PHRBench: A Behavioral Evaluation of Post-Hallucination Reasoning in LLMs

cs.CL cs.AI ·2026-10-07 ·Linghao Meng, Feng He, Xuan Yang, Junyuan Mao, Pinze Ren +3 more
benchmark
Abstract

Hallucinated information can propagate through multi-stage LLM systems and become part of the context for subsequent reasoning. Existing studies of post-hallucination reasoning (PHR) mainly characterize changes in final outcomes and aggregate reasoning dynamics, leaving how models resolve hallucinated premises at the response level insufficiently understood. In this work, we introduce PHRBench, a controlled benchmark for behaviorally structured PHR across four domains and 18 large language models. PHRBench characterizes each reasoning trajectory independently of final-answer correctness through Hallucination Compliance, Hallucination Avoidance, and Heuristic Correction, and defines an insightful trajectory as successful correction that ultimately reaches the correct answer. Across 4820 controlled instances, we find that successful recovery remains relatively rare and is associated with more frequent belief updates along the reasoning trajectory. We further find that properties of the hallucinated prompt contain substantial predictive signal for successful recovery, with a lightweight predictor achieving an AUROC of 0.847. These findings provide a behavioral view of post-hallucination reasoning, characterizing how LLMs resolve erroneous context and when successful recovery is likely to occur.

0.14

DisParQ: Self-Supervised Part Concepts for Interpretable Vision Foundation Models

cs.CV cs.AIcs.LG ·2026-10-07 ·Adam Pardyl, Siddhartha Gairola, Sukrut Rao, Adam Wróbel, Bartosz Zieliński +2 more
Abstract · Under review

Concept-based vision models represent images through an intermediate layer of human-inspectable concepts, so what a model relies on can be traced to those concepts. However, those models are often limited to fixed categories or depend on language to define their concepts. We introduce DisParQ (Discrete Parts with Quantized attributes), a method that learns spatially grounded, discrete concept representations from a powerful frozen vision-only self-supervised backbone. It requires no class labels and no language supervision. Each image patch is assigned to exactly one concept from a learnable prototype dictionary, and only a sparse subset of concepts may activate per image. To capture how each concept varies across images (e.g., the type of a "wheel"), we learn continuous residuals alongside the concepts and then quantize them into discrete attributes. A spatial decoder reconstructs the backbone's representation from the concepts and attributes alone, so successful reconstruction means that the discrete representation preserves the backbone's information. We evaluate DisParQ across seven datasets, from general recognition (ImageNet, PartImageNet, Places) to fine-grained benchmarks (CUB, Cars, Dogs, Flowers). We show that DisParQ closely matches its frozen DINOv2 teacher on ImageNet linear probing (83.2% top-1), achieves higher concept consistency than language-aligned models, remains competitive on fine-grained recognition, and enables cross-category part-based retrieval.

0.14

A multi-scenario EEG dataset for auditory attention decoding in naturalistic multi-talker environments

cs.SD ·2026-10-07 ·Shu Peng, Rui Liu, Yufei Zhang, Wenlong You, Zhige Chen +5 more
benchmarkdatasetcode/data
Abstract

Understanding how the brain selectively follows relevant speech amid competing voices is a central challenge in auditory neuroscience and a key step toward neuro-steered hearing technologies. However, most open-source Electroencephalography (EEG) datasets for Auditory Attention Decoding (AAD) use idealized single-competing-talker paradigms that oversimplify the acoustic, spatial, and semantic structure of everyday communication. To capture this ecological complexity, we introduce the SoundBubble-EEG dataset: a high-density 128-channel EEG resource comprising more than 25 hours of recordings from 30 participants. The paradigm requires listeners to selectively attend to a dynamic target speaker group, a designated "sound bubble", amid competing multi-speaker distractor bubbles across three realistic scenarios: a restaurant, a home TV viewing, and a meeting discussion. By bridging the gap between constrained laboratory protocols and real-world auditory scenes, this dataset enables investigations of multi-talker speech comprehension, neural speech tracking, and cross-scenario generalization. It also provides a benchmark for AAD algorithms under realistic acoustic and semantic variability and may support auditory neuroscience and the development of neuro-steered hearing technologies.

0.13

Omni-Diffusion-Distill: Few-Step Distillation of Unified Multimodal Diffusion Large Language Models

cs.CV cs.LG ·2026-10-07 ·Hong Huang, Chenhongyi Yang, Junzhe Sun, Animesh Sinha, Wuyang Chen +1 more
Abstract

Unified multimodal diffusion large language models (dLLMs) offer a single architecture for both image generation and multimodal understanding, but their iterative decoding requires tens to hundreds of forward passes. Existing few-step distillation methods largely focus on either image generation or text generation, making it unclear how to compress a fully discrete multimodal dLLM into a single efficient student while preserving both generation and understanding. We introduce Omni-Diffusion-Distill, a unified two-stage distillation framework that retains strong generation and understanding capabilities while substantially reducing the inference cost of a unified multimodal dLLM. Omni-Diffusion-Distill aligns the distillation of both generation and understanding, for both images and text, in the discrete token space. In the first stage, the student is trained to skip decoding steps by replaying cached teacher trajectories, and in the second stage the student is refined on intermediate states along its own rollouts. We further remedy two sources of degradation in unified distillation with a pairwise collision penalty that reduces repetition under parallel text decoding, and entropy-matched guidance that prevents entropy collapse caused by fitting the sharpened teacher distribution in image generation. Omni-Diffusion-Distill achieves state-of-the-art trade-offs between decoding efficiency and generation and understanding performance for multimodal dLLMs, reducing image generation from 128 to 8 decoding steps and multimodal understanding from 512 to 64, giving 18.2x and 21.2x wall-clock speedups. Under these budgets, it scores 0.828 on GenEval and 83.0 on DPG-Bench for text-to-image generation, while reaching GPT judge scores of 20.0 on MM-Vet and 57.2 on COCO captioning (twice the teacher's 28.4 at the same steps) for multimodal understanding.

0.13

CARES: A Controlled Synthetic Benchmark of Speaker Reactions to Sound

cs.SD cs.LG ·2026-10-07 ·Marcel Gibier, Thomas Thebaud, Olivier Boëffard, Jean-François Bonastre
benchmark
Abstract · Submitted to ICASSP 2027

Automatic audio scene description turns a recording into a text account of a situation. One difficulty is deciding which elements of the audio should be kept, since a description cannot include them all. Annotators disagree about this, making a ground truth hard to obtain. In this work, we first define the ground truth, then generate the data. We focus on audio events and define sound salience with a simple rule: a sound is salient when a speaker audibly reacts to it. For scale and variety, a controlled set of scenarios fixes the ground truth, and a language model writes the dialogues. The resulting corpus, CARES, contains 10,000 two-speaker scenes. We then benchmark six audio-language models on three tasks: identifying the scene, tagging the sounds present, and classifying reactions. We show that these models hear the sounds but miss how the speakers react to them.

0.13

InscriptionOCR: A Dataset and Method for Understanding Inscriptions

cs.CV ·2026-10-07 ·Jaidev Sanjay Khalane, Akbar Ali, V. N. Prabhakar, Shanmuganathan Raman
Abstract

Ancient script image restoration is a fundamental problem in computer vision, as it directly affects the reliable analysis and interpretation of historical documents and inscriptions. Ashokan Brahmi is an ancient script extensively used during the reign of Emperor Ashoka in the 3rd century BC, primarily for inscriptions in Prakrit. These inscriptions, including major and minor rock and pillar edicts, constitute a valuable yet largely unexplored source of data for computational analysis. The degraded nature of inscription imagery and the lack of standardized digital resources pose significant challenges for automated processing. We present an end-to-end AI-based framework for understanding ancient inscriptions that encompasses image enhancement, optical character recognition (OCR), transliteration, and neural machine translation (NMT). The proposed pipeline processes low-quality images captured directly from stone inscriptions, performs image restoration and Brahmi script character recognition, maps the recognized characters to the Roman script, and finally translates the resulting Prakrit text into English. We also introduce two new datasets: (i) InscriptionOCR Dataset: the largest publicly usable digital OCR dataset for Brahmi script to date, consisting of over 200,000 character images across about 600 classes, and (ii) a bilingual Prakrit-English parallel corpus comprising over 2,000 sentence pairs for NMT. We believe that the proposed framework and datasets will facilitate future research in ancient script analysis, low-resource OCR, and digital epigraphy.

0.13

ActiveLang: Active Open-Vocabulary 3D Mapping with Semantic-Uncertainty-Guided Exploration

cs.CV cs.RO ·2026-10-07 ·Liyan Chen, Hairong Yin, Huangying Zhan, Yi Xu, Raymond A. Yeh +1 more
Abstract

As robots increasingly assist humans with diverse tasks, they need both geometric and semantic understanding of their surroundings. Moreover, robots often operate in unfamiliar environments and take on new tasks without knowing the relevant concepts ahead of time. This motivates language-annotated 3D maps that support open-vocabulary scene understanding and human-robot interaction. We introduce ActiveLang, an autonomous system for active open-vocabulary 3D mapping with semantic-uncertainty-guided exploration. ActiveLang performs online language-feature adaptation on a compact dual-Gaussian representation to jointly reconstruct scene geometry, appearance, and open-vocabulary semantics with modest memory overhead. Its planner efficiently selects informative viewpoints, enabling effective mapping with fewer observations and lower computational cost. Experiments on Replica and ScanNet++ demonstrate substantial improvements in 2D and 3D open-vocabulary segmentation over both online and offline baselines, highlighting that actively exploring scenes builds language-annotated 3D maps more efficiently.

0.13

Listen-to-Reason: Listen with Experts, Retrieve over a Graph, Reason with LLMs

cs.SD ·2026-10-07 ·Pooneh Mousavi, Mirco Ravanelli, Cem Subakan
Abstract

Large audio-language models (LALMs) fuse an audio encoder into a large language model (LLM) through multi-stage training. This coupling means that a new domain or a stronger LLM requires retraining, and their answers cannot be traced to what the model heard: a chain-of-thought is a post-hoc account. We propose Listen-to-Reason (L2R), an interpretable-by-design pipeline that passes audio to the LLM through an explicit, human-readable tree: small heads on frozen expert encoders map each chunk of a clip to semantically meaningful nodes on the tree (for speech, music and environmental sound), and a frozen text-only LLM answers from these nodes and an ASR transcript without hearing the clip. Every answer can therefore be traced to the nodes and transcript it read, and the nodes are causal: replacing the deciding node with a distractor overturns 78% of correct answers on SAKURA. With a 7B reader, L2R outperforms all LALMs we compare against on SAKURA and trails them by 6-12 points on MMAU and MMAR, despite training about 1,400x fewer parameters on orders of magnitude less audio data. However, because any LLM can serve as the reader, we show that a stronger reader narrows this gap without retraining any audio component. A new domain is added with one small head: with five labelled clips per species, it outperforms QLoRA fine-tuning of an LALM on the same clips by 13-26 points.

0.13

DynStream: Online Streaming 4D Gaussian Reconstruction of Dynamic Worlds from Unposed Video

cs.CV ·2026-10-07 ·Dingwei Xian, Xiaoyu Zhou, Yajiao Xiong, Yongtao Wang, Ming-Hsuan Yang
Abstract

Online reconstruction of dynamic 4D scenes from long, unposed streaming videos requires both continuous processing and photorealistic rendering, which existing methods struggle to achieve simultaneously. Existing feed-forward Gaussian methods are restricted to offline processing, whereas online point-cloud approaches struggle to maintain dense geometry and high-fidelity rendering. We present DynStream, a framework for streaming 4D Gaussian reconstruction from long, unposed videos. Given a continuous video stream, DynStream reconstructs the scene within local temporal windows and incrementally aligns and fuses these local reconstructions into a globally consistent scene, enabling online 4D reconstruction without per-scene optimization. By jointly enforcing cross-window geometric consistency and modeling time-varying scene content, DynStream supports efficient reconstruction and photorealistic rendering over extended video streams. Experiments demonstrate that DynStream enables high-fidelity online dynamic reconstruction and rendering from long video streams, achieving state-of-the-art performance across diverse dynamic indoor and outdoor scenes.

0.13

Beyond Outcome Rewards: Constructing and Assigning Retrieval Credit for Search Agents

cs.CL cs.AIcs.LG ·2026-10-07 ·Wenyu Huang, Xinyu Hou, Pavlos Vougiouklis, Ruofei Lai, Jeff Z. Pan
Abstract

Search agents enable Large Language Models (LLMs) to iteratively retrieve and use information for complex multi-hop questions. Reinforcement Learning with Verifiable Rewards (RLVR) offers a promising approach for post-training such agents, but its reliance on sparse, outcome-based supervision can make credit assignment difficult and limit learning efficiency. In this paper, we systematically investigate how intermediate supervision can improve reinforcement learning for search agents. We study a range of reward-shaping and credit-assignment strategies that provide learning signals from intermediate retrieval steps. Building on these insights, we develop a training framework that combines intermediate signals with final outcome rewards to improve learning from multi-step search trajectories. Experiments across multiple benchmarks under matched training conditions demonstrate improvements in aggregate search-agent performance and show that both the choice of intermediate signal and where its credit is assigned affect training behaviour. These findings show that reward design and credit assignment are important design dimensions for training effective search agents.

0.12

A Lightweight Hybrid Framework for Acoustic Echo Cancellation and Suppression via Constrained Recursive Training

eess.AS ·2026-10-07 ·Huawei Zhang, Rilin Chen, Hao Zhang, Meng Yu, Dong Yu
Abstract · Accepted by the 19th International Workshop on Acoustic Signal Enhancement (IWAENC 2026)

We propose a hybrid framework that combines deep learning with adaptive filtering for acoustic echo cancellation (AEC) and suppression. It integrates a multi-output neural network to jointly update filter parameters and perform post-processing. A learnable variant of the normalized least mean squares (NLMS) filter is introduced, incorporating a variable step size and a variable transition factor that are recursively updated for dynamic adaptation. In addition, an enhanced near-end speech estimate is generated from the filter output to suppress the residual echoes. To ensure training stability, we propose a constrained recursive training strategy that aligns with the adaptive nature of the filter, where an upper-bound constraint is introduced on the filter output. For efficiency, the framework adopts a lightweight implementation with only 253 k parameters. Experimental results show that the framework outperforms the baselines in terms of both signal quality and speech intelligibility.

0.12

Trimodal Context-Aware Micro-Expression Recognition in Real-World Conversations: A Dataset and a Hierarchical Fusion Network

cs.CV cs.MM ·2026-10-07 ·Junbo Wang, Yan Zhao, Shigang Wang, Jian Wei, Qiming Zhang +1 more
Abstract

Recognizing micro-expressions in real-world conversations is challenging, as these facial movements are brief and faint, and their interpretations are strongly dependent on the contextual circumstances of their occurrence. Most existing methods are de-contextualized, as they analyze isolated video clips while disregarding the semantic and acoustic cues in the surrounding dialogue that help explain the emergence of an expression. Decades of psychological studies have shown that social context exerts influences on facial behavior over two key dimensions: how an expression is interpreted and how strongly it is generated to begin with. However, existing micro-expression research rarely incorporates such contextual factors into its models. As an example, provocative utterances from an interlocutor may elicit suppressed negative affective reactions, which isolated silent clips are unable to account for. We propose TriCoME, a trimodal dataset consisting of visual, audio and text data sourced from real-world conversations. In this dataset, each micro-expression is paired with its concurrent utterance and preceding context. We further annotate the context based on conversational role to denote two categories: external stimuli from an interlocutor's utterance and internal background from the subject's own utterance. On this basis, we propose the Hierarchical Focusing Fusion Network (HFFNet), which aligns acoustic and semantic streams at fine granularity, compresses long dialogue sequences, modulates attention toward conversational roles via a context-type embedding, and routes each sample between visual-only and multimodal decisions based on visual confidence to prevent noisy context from degrading recognition performance.

0.12

LLM Persuasion Is in the Eye of the Evaluation

cs.CL cs.AIcs.CY ·2026-10-07 ·Kamile Dementaviciute, Julija Vaitonyte, Tijl De Bie
Abstract

Large language models (LLMs) have already been shown to match or exceed human experts in persuasion. While their persuasive capabilities hold promise for beneficial uses such as education and health communication, they can also be used to manipulate and misinform, making their evaluation a growing priority for developers and regulators. That evaluation, however, remains fragmented: studies differ in what they treat as persuasion, and broad claims often rest on narrow, situation-specific assessments. Automated methods, often modelled on human studies, offer a way to compare such assessments directly, as they can be run on the same models at scale and can include high-risk forms of persuasion that would be difficult or unethical to test on people. In this study, we adapt nine published automated methods to a shared setup, run them on the same fifteen LLMs, and ask whether their rankings agree and why. We find that the methods agree only weakly (mean Spearman $ρ= 0.25$). Our analyses point to two contributing factors. Models that refuse some tasks but not others, directly or indirectly, lower agreement by about a quarter, and these refusals fall mostly on manipulation tasks. General capability also plays a part: most rational persuasion (non-manipulative) methods track it, whereas most manipulation methods do not. Together, these findings suggest that agreement depends more on the task a method sets than on how it scores persuasion, although this pattern is only indicative given the eight methods available for analysis. More broadly, our results suggest that persuasion scores combine a model's ability to persuade with its willingness to do so. A single score is therefore informative about its own setting, but says little about a model's persuasiveness across tasks.

0.11

Steerspeech: Activation Steering For Emotion Control In Generated Speech

cs.SD cs.LGeess.AS ·2026-10-07 ·Afsara Benazir, Darius Pétermann, Felix Xiaozhu Lin, Salar Rahili
Abstract · Under review at IEEE ICASSP 2027. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without no…

Pretrained text-to-speech (TTS) models can generate expressive speech, but reliable inference-time emotion control remains challenging: prompts and reference audio offer coarse, inconsistent control, whereas specialized conditioning and model adaptation require costly training. We present SteerSpeech, a lightweight activation-steering framework that controls emotion by injecting steering vectors into hidden activations. For each target emotion we train a lightweight low-rank transform, using a multi-expert objective that encourages monotonic emotion control while preserving speaker identity and linguistic content, constraining steering drift, and keeping the TTS backbone frozen. To optimize through discrete speech tokens, we introduce a two-pass generation-and-replay pipeline using a straight-through estimator to backpropagate expert supervision through sampled tokens. At inference, a target-emotion steering direction is optimized with its respective transform and injected into the base TTS model. Objective and subjective evaluations with Qwen3-TTS across seen, unseen, and accented speakers show stronger continuous emotion control with limited speaker and content degradation. SteerSpeech achieves 1.08x-7.12x baseline target-emotion scores and for a representative emotion subjectively, it receives 78.1%-96.8% intensity preference and 1.43x-1.46x speaker-identity preservation at high steering strengths.

0.07

InterView-C: A Synchronized Multimodal Corpus of VR Avatar-Mediated Survey Interviews

cs.CL ·2026-10-07 ·Patrick Schrottenbacher, Leon Hammerla, Lydia Kleine, Doris Stingl, Alexander Mehler
Abstract

We present InterView-C, a German multimodal corpus of 27 survey interviews conducted entirely in virtual reality, with both interlocutors represented by avatars. The corpus aligns spoken interaction with synchronized behavioral data, including gaze, head and body movement, facial behavior, hand and finger tracking. Its reference transcripts and linguistic annotations provide a reliable interface between this multimodal spoken interaction and predominantly text-based NLP methods. This interface is important because automatically transcribing speech can distort linguistically relevant information, while downstream models trained on existing resources may additionally face transfer challenges when applied to transcribed spoken data. InterView-C therefore provides word-timed and manually post-edited verbatim transcripts for all 54 recordings, interview-item timings, questionnaire responses and negation cue and scope annotations for 1,422 sentences, 1,398 of them doubly annotated (α=0.87 for cues; α=0.81 for scopes). We demonstrate both challenges empirically: nine open-weight ASR systems disproportionately misrecognize short closed answers and number words, while negation models trained on existing corpora show lower and highly variable performance on our transcribed interviews than a model trained on the InterView-C annotations. InterView-C thus enables linguistic analyses of spoken interaction while retaining their alignment with rich multimodal behavior.

0.07

PARC-Loc: Text-to-Point-Cloud Localization with Partial Assignment and Relational Consistency

cs.CV cs.AIcs.CL ·2026-10-07 ·Shengkai Ma, Zhenyu Hou, Weihua Cao
Abstract

Text-to-point-cloud localization estimates a position in a city-scale 3D map from descriptions of surrounding objects. Existing coarse-to-fine methods retrieve submaps using aggregate learned compatibility and then localize within a selected submap. However, repetitive or similar urban objects can inflate the embedding similarity between the query and multiple submaps, even when the instance layout within a submap violates the query description. Meanwhile, query-relevant instances often span submap boundaries, leaving the retrieved submap with incomplete contextual evidence. We term these failure modes layout-inconsistent aliasing and boundary evidence incompleteness, respectively. To address them, we propose PARC-Loc, a coarse-to-fine localization framework built on Partial Assignment with Relational Consistency (PARC). PARC jointly models hint-object compatibility and pairwise spatial relations, allowing unmatched elements while favoring assignments consistent with the queried layout. At the coarse stage, its candidate-level assessment complements neural similarity for layout-consistent submap selection. At the fine stage, the context is expanded with query-relevant instances from adjacent submaps, while PARC yields object-level matching weights that guide cross-modal attention. Extensive experiments on KITTI360Pose and CityLoc show that PARC-Loc outperforms conventional coarse-to-fine baselines. On KITTI360Pose, our method improves Top-1 localization recall at 5 m from 0.50 to 0.67, achieving a 34% relative gain over the strongest baseline.

0.05

EASE: Entropy-Adaptive Distribution Shaping for Evading AI-generated Text Detectors

cs.CL ·2026-10-07 ·Jicheng Zhou, Kahim Wong, Jialong Wang, Jiantao Zhou
Abstract

AI-generated text (AIGT) detection can be sensitive to the decoding choices of the source large language model (LLM). We observe that perturbing next-token logits or adjusting sampling temperature can reduce detection performance, providing a clear signal of detector vulnerability to decoding-time distribution changes. Building on this observation, we propose EASE (Entropy-Adaptive Distribution Shaping for Evasion), a training-free and detector-agnostic framework for evading AIGT detectors. EASE computes predictive entropy directly from the source LLM's next-token distribution and uses it to adapt both logit perturbation and sampling temperature, without detector feedback or model fine-tuning. Experiments across three source LLMs and multiple detectors demonstrate consistent reductions in detection performance, with negligible degradation in text quality and negligible inference overhead.

0.04

TIRA: Tumor Immune Representation Adaptation for Zero-Shot Cross-Cancer MSI and TMB Prediction

cs.CV ·2026-10-07 ·Dasari Naga Raju
Abstract

Microsatellite instability-high (MSI-H) and high tumor mutational burden (TMB-H) are clinically relevant biomarkers, yet their histopathological prediction remains challenging when models are transferred across morphologically distinct cancer types. Immune-associated spatial patterns can persist across cancers despite these morphological differences, but foundation-model-based predictors trained on a single cancer do not explicitly use this information, limiting cross-cancer generalization. To address this limitation, we propose TIRA (Tumor Immune Representation Adaptation), a target-free framework that refines frozen foundation-model representations using spatial immune topology, without requiring target-domain data during model development or test-time adaptation. TIRA uses a topology-supervised biology representation to condition tile-level attention while pooling only morphological features for joint MSI and TMB prediction. We train TIRA on TCGA-COAD+READ and evaluate it zero-shot on CPTAC-COAD, TCGA-STAD, TCGA-UCEC, and CPTAC-UCEC, covering cross-site, cross-cancer, and combined cross-cancer-site distribution shifts under UNI2, CONCH, and Virchow2. With UNI2, TIRA improved zero-shot AUROC on TCGA-STAD from 0.633 to 0.766 for MSI and from 0.651 to 0.772 for TMB. Source-derived spatial immune topology improved the cross-cancer robustness of frozen pathology foundation-model representations.

0.04

Source-Directed Trajectory Perturbation at First-Order Cost for Domain Generalization in Speech Deepfake Detection

eess.AS ·2026-10-07 ·Siqing Qin, Kong Aik Lee, Youzhi Tu
Abstract

Speech deepfake detectors often lose accuracy when the distribution of the test data differs from that of the training data. Meta-learning for domain generalization (MLDG) shows promise by simulating domain shifts with episodic meta-train and meta-test splits. However, the MLDG meta-objective considers only a single clean adaptation trajectory and does not account for the local loss landscape around its endpoints. To address this limitation, we first propose an explicit worst-case MLDG variant, dubbed WC-MLDG-4P, which perturbs both the meta-train and meta-test states but requires four gradient evaluations per episode. We then introduce WC-MLDG-2P, a source-directed alternative to the explicit robust objective. It shifts the clean MLDG endpoint toward a locally higher source-loss state and evaluates the meta-test gradient there, retaining the two-evaluation cost of the first-order MLDG. A first-order expansion relates the resulting gradient change to meta-test directional curvature along the source gradient without explicitly computing a Hessian. Relative to MLDG, WC-MLDG-2P achieves relative mean-EER reductions of 23.4% with XLSR-AASIST and 7.6% with XLSR-Conformer-TCM, respectively.

0.02

Security & Privacy 41

Understanding and Mitigating Token-Pruning-Induced Vulnerabilities in VLMs

cs.CR cs.LG ·2026-10-07 ·Shuailong Wang, Xinyu Lyu, Shengming Yuan, Jingkuan Song, Heng Tao Shen +1 more
top venue
Abstract · Accepted at ICML 2026. 16 pages

Token-Pruning accelerates Vision-Language Models by removing redundant visual tokens, yet its safety implications remain underexplored. In this work, we present the first comprehensive safety evaluation of Token-Pruning mechanisms and find that: most pruning strategies significantly degrade safety as pruning ratios increase, whereas Query-based Compression shows the opposite, with extreme pruning (up to 99.8%), unexpectedly improves model safety. This sharp contrast prompts a key question: How do different Token-Pruning strategies reshape model safety behavior, and is it possible to enhance safety without sacrificing acceleration? To answer this, we identify an unrecognized mechanism, termed Pruning-Induced Malicious Amplification, where removal of background tokens triggers a side effect: forcing the model's attention to collapse onto a few retained malicious anchors within the foreground, inadvertently amplifying their toxic semantics under jailbreak. To address that, we propose an inference-time and plug-and-play Safety-Aware Pruning (SAP) mechanism that counteracts such dominance via three steps: (1) identifying malicious anchors, (2) restoring pruned benign tokens, and (3) reallocating excessive attention from malicious anchors to benign tokens. Extensive experiments across three safety and four utility benchmarks demonstrate that SAP mitigates pruning-induced vulnerabilities, i.e., reducing ASR by up to 62%, without compromising efficiency or utility.

0.64

Sensitive-Topic Leakage Through LLM Routing Metadata: Measurement and Mitigation

cs.CR cs.CL ·2026-10-07 ·Teng-Ruei Chen
Abstract · 20 pages, 5 figures, 9 tables

LLM routers pick a cheap or expensive model per request by its content, and many gateways and some cloud platforms can log that choice with content logging off. We measure this privacy channel beyond token counts, accounting for noisy labels and repeated prompts. We run pre-registered studies on 1.7 million real requests (WildChat-1M, LMSYS-Chat-1M) with two cost/quality routers and a domain router, survey eleven systems' logging, and test post-processing defenses. At matched length, the shift's direction depends on category and router. For RouteLLM at the 50% operating point, harassment and self-harm requests reach the strong model 19 points less often than comparable ones on prompts unseen in exploration, medical requests (exploratory: LLM labels failed their gate) 31 points less often on distinct prompts (both post hoc), and sexual requests 10 points more often (secondary); the other router's four are negative. Twenty RouteLLM decisions separate frequent medical askers with AUC 0.71, exploratory and below the pre-registered primary endpoint's 0.75 (domain router: 0.92, an upper estimate). Per-category length-matched parity with accurate labels removes the gap on real traffic, costing at most 0.2 accuracy points on RouterBench (post hoc), where routers' gaps on sensitive subjects (13-42 points, pre-registered) exceed those of an oracle routing by realized accuracy gain (1-11, post hoc). Per-conversation stickiness, per-user budget bands, and pooled parity fail, the last as categories' shifts differ in size or sign. A post hoc exact per-user rate hides only even-prefix strong counts and forfeits most self-assessed routing value; it preserves odd-position decisions, from which a post hoc log attack reaches AUC 0.73 after 20 RouteLLM requests (exploratory).

0.64

CYBERFORT: A Compliance-Chain Platform Operationalising the Cyber Resilience Act for SMEs

cs.CR cs.CYcs.SE ·2026-10-07 ·Nikolaos Kekatos, Georgios Koutidis, Mihaela Curcă, Angelos Fountoulakis, Marios Ioannou +6 more
publishedcode/data
Abstract · 7 pages, 3 figures, 2 tables. Accepted at the 2026 IEEE International Conference on Cyber Security and Resilience (IEEE CSR 2026)

The EU Cyber Resilience Act (CRA) turns product cybersecurity into a lifecycle compliance obligation for manufacturers, importers, distributors, and integrators of products with digital elements on the EU market, a load that falls largely on small and medium-sized enterprises (SMEs) that rarely have dedicated governance, risk, and compliance (GRC) capacity. We present CYBERFORT, an open-source CRA-first compliance platform developed under the EU Digital Europe Programme and one of twelve projects in the EU CRA cluster. CYBERFORT operationalises the CRA through a guided scope self-assessment, a question bank tied to Annex I and the vulnerability-handling obligations, and a compliance-checking engine that links every answer to controls, policies, and machine-attested evidence, reusing ISO/IEC 27001, NIS2, and GDPR controls only where they coincide with CRA obligations. Its central contribution is the compliance chain, a traceable structure linking each product risk through its controls and policies to the CRA obligations it satisfies, and onward through evidence to the technical-documentation file and EU declaration of conformity, so that every operational gap is traceable and can be closed before market placement. Deployed at https://access.cyber-fort.eu/ for a first cohort of 43 organisations, the platform is presented with the engineering behind the chain, measured results from a completed end-to-end case study on a SIEM/XDR product with AI-driven remediation spanning the CRA obligation chapters, and the controlled effort study that remains in progress.

0.64

Constrained-Action AI Remediation for SIEM/XDR via a NeMo-Guardrails Proxy

cs.CR cs.AIcs.CL ·2026-10-07 ·Georgios Koutidis, Nikolaos Kekatos, Tom Nianios, Alexios Lekidis
publishedhighly cited
Abstract · 7 pages, 3 figures, 4 tables. Accepted at the 2026 IEEE International Conference on Cyber Security and Resilience (IEEE CSR 2026)

Security Operations Centers (SOCs) for information technology and operational technology share one incident-response problem: a flood of correlated alerts and too few analysts. Large Language Models (LLMs) are increasingly proposed as reasoning engines that triage alerts and, in autonomous deployments, issue commands that block IPs, kill processes, or quarantine files on production hosts. This coupling introduces a new risk: a single adversarial alert can become a remote code path through the LLM's reasoning, leading it to recommend an action the SOC then executes. We present a constrained-action architecture with two coordinated layers: (i) a SIEM/XDR control plane that grounds remediation in correlated host events and confines the LLM's output to a closed intent vocabulary whose templated commands are executed by thin endpoint agents, backstopped by an argument validator; and (ii) a NeMo-Guardrails proxy that wraps the SOC-analyst LLM with input- and output-rail policies, evaluated out-of-the-box against a SOC-specific adversarial corpus we release. The stock proxy lifts injection recall from 25.0% to 94.5% at a 0.1% false-positive rate, and a live red-team exercise confirms that the closed intent vocabulary and argument validator contain the observed LLM failure modes before any command crosses the trust boundary. As an architectural fit (not yet a measured operational-technology deployment), the constrained-action property suits critical-infrastructure settings where a wrong remediation has physical, not merely operational, consequences. The loop is best run human-in-the-loop or delayed: the measured rail latency keeps inline control out of scope.

0.58

Hybrid Hierarchical Runtime Verification for Edge-IoT Security: Combining MonPoly and RTLola

cs.CR cs.DCcs.LO ·2026-10-07 ·Nikolaos Kekatos, Marinelio Chintri, Panagiotis Katsaros, Alexios Lekidis, Tom Nianios +3 more
publishedhighly cited
Abstract · 19 pages, 2 figures, 7 tables. Published in Availability, Reliability and Security (ARES 2026), EU Projects Symposium Workshops, Springer LN…

Security monitoring of edge-IoT fleets faces three structural challenges. (i) A per-node monitor is cheap but cannot see attacks that coordinate across devices. (ii) A cloud monitor sees the full fleet but pays for that view in bandwidth. (iii) Even at the cloud, a monitor built on a single RV engine can be fooled by an attacker who compromises a device, raises one malicious request, and then goes silent: once the events stop, an event-triggered monitor has nothing left to evaluate. We propose a three-layer hierarchical runtime-verification framework that addresses all three. The edge layer classifies events as they happen, the gateway layer aggregates short windows of per-device behaviour, and the cloud layer runs two complementary RV engines. MonPoly handles first-order temporal correlation over the merged alert stream: coordinated overflow (which genuinely quantifies across devices) plus per-device multi-vector APT, escalation, and persistent-campaign patterns. RTLola handles a time-triggered silent-node property that an event-triggered engine cannot detect within a bounded delay under fleet silence. We evaluate the framework on a 15-actor Docker testbed covering eight attack profiles plus a silent-bypass scenario. In the controlled labelled testbed, every device-attributable incident the framework raises names an attacker-labelled device, and the RTLola tier catches silent-bypass attempts the event-triggered tier misses. Per-event monitoring stays in the microsecond range at the edge and gateway, with low end-to-end alert-to-incident latency at the cloud.

0.58

SLDR: Defending Against Malicious Fine-tuning via Selective Layers Recovery and Dynamic Routing

cs.CR cs.AI ·2026-10-07 ·Hui Zhang, Yachao Yuan, Jiayun Wang, Yuanzhuo Li, Hongtao Wang +1 more
top venuecode/data
Abstract · Accepted at NeurIPS 2026

Fine-tuning-as-a-service enables users to adapt aligned large language models (LLMs) to specialized tasks, but malicious fine-tuning can erode refusal behavior while preserving task performance on legitimate inputs. We revisit recent layer-wise safety diagnostics and find that safety sensitivity is signed: scaling different layers can strengthen refusal, weaken it, or have little effect. Motivated by this observation, we propose SLDR, a post-fine-tuning defense based on Selective Layers Recovery and Dynamic Routing. SLDR trains a LoRA recovery adapter only on the layers with the maximum and minimum sensitivity scores in the signed spectrum, and uses representation-based dynamic routing inference to activate the adapter only for malicious queries. Across four model architectures, five downstream tasks, and four harmful benchmarks, SLDR substantially reduces harmful outputs while preserving downstream utility. On Llama3.1/SST2, SLDR reduces the average harmful score from 11.54 to 0.08 while maintaining downstream accuracy, and the harmful score remains near zero under poisoning ratios up to 0.9. The code is available at https://github.com/Stardust457/SLDR.

0.56

Formal Runtime Verification for Tool-Using LLM Agents: An Offline Same-Benchmark Study on AgentDojo and STAC

cs.CR ·2026-10-07 ·Nikolaos Kekatos, Stylianos Basagiannis, Marinelio Chintri, Alexios Lekidis, Tom Nianios +3 more
publishedbenchmark
Abstract · 14 pages, 1 figure, 13 tables. Accepted at the 8th Workshop on CPS&IoT Security and Privacy (CPSIoTSec '26), co-located with ACM CCS 2026, T…

Guardrails for tool-using LLM agents are usually application-specific rules, which makes multi-step, data-dependent safety policies hard to specify, audit and reuse. As a declarative alternative, we evaluate metric first-order temporal logic (MFOTL), replaying the recorded trajectories that AgentDojo, STAC and R-Judge already ship through the unmodified MonPoly monitor, offline and without running an agent. On these corpora, five generic obligations flag 71.8% of STAC attack chains and 70.1% of successful AgentDojo attacks, but also fire on 29.3% of benign runs. This imprecision stems from the corpora rather than the logic: they rarely record approvals and never record timestamps, so history-dependent obligations reduce to detecting risky action types. Where the trace does carry relational context, provenance-aware policies discriminate better; that context, however, is itself attackable, and one planted line defeats a naive provenance check on 94-99% of the runs it would otherwise flag. Binding provenance to the lookup that produced it closes this evasion at no cost in detection or benign firing. Taken together, these results show that formal temporal monitoring adds value exactly when the trace exposes trustworthy history. We therefore quantify how far current benchmarks are from that point and propose a twelve-field enforcement-ready trace schema.

0.55

Hierarchical Security Monitoring for Edge-IoT: A Formal Methods Approach

cs.CR cs.LOcs.SE ·2026-10-07 ·Nikolaos Kekatos, Marinelio Chintri, Panagiotis Katsaros, Alexios Lekidis, Tom Nianios +2 more
publishedhighly cited
Abstract · 7 pages, 1 figure, 2 tables. Accepted at the 2026 IEEE International Conference on Cyber Security and Resilience (IEEE CSR 2026)

Cyber resiliency in edge-IoT deployments is fundamentally an economic problem: detection must keep critical processes operating under attack, but defender resources (compute, bandwidth, operator attention) are bounded. Centralised cloud monitoring offers expressive cross-device detection at prohibitive bandwidth cost; purely edge-local monitoring is cheap but blind to coordinated multi-device attacks where the asymmetric balance favours the attacker. We propose a lightweight hierarchical security-monitoring framework, built on formal runtime-verification methods, that occupies the practical middle ground at quantified cost. Each edge device runs a lightweight TeSSLa stream specification (size, payload validity, rate, and timestamp-drift predicates) that emits a four-valued verdict per aggregation window at sub-microsecond per-event cost; the gateway runs a parametric first-order MonPoly monitor over the per-device verdict streams at microsecond-scale per-verdict cost. The edge-to-gateway uplink carries roughly one Boolean per aggregation window per node, orders of magnitude smaller than the raw packet stream. The gateway tier detects coordinated attack patterns that no single-node monitor can see, shifting the asymmetric cost balance toward the defender. Every alert carries a witness set naming the device, the monitor tier, and the predicate that fired, providing an auditable record of the decision. We evaluate the framework on a container-host testbed spanning nominal and attacker nodes across four attack classes (buffer overflow, time spoofing, denial-of-service, and mixed advanced-persistent-threat patterns), and describe the edge- and gateway-tier specifications together with the cost-versus-coverage trade-off as monitor levels are added.

0.54

When AI Finds Hidden Messages, Does It Report?

cs.CR cs.AI ·2026-10-07 ·William Guey, Rashik Jahangir, Pierrick Bougault, Vitor D. de Moura, Wei Zhang +1 more
code/data
Abstract · 2 figures, 8 tables. Data and code (v1.0.0): https://github.com/williamguey/ai-hidden-message-reporting

When an assistant encounters a message for another AI, does it tell its user? Four fixed model-provider deployments perform simulated source tasks in 1,280 ordinary-note and 128 enhanced-note sessions. Harmless and harmful messages have matched plaintext and ROT13 versions, with no-message controls. Observers receive no decoder or decoded meaning; a requested reference code incentivizes inspection. Asking for reports increases rule-detected notifications identifying another AI as recipient by 53.1 percentage points for harmless ROT13 messages and 54.7 for harmful ones. This is a joint inspection, recognition, and notification effect; missing-response bounds are 38.3--77.3 and 36.7--78.1 points. Model-based trace checks identify eleven ordinary plaintext cases where agents interpret the message but do not notify their user. Seven encoded omissions are verified with enhanced notes; ordinary encoded omissions remain unverified. Seven simulated filename disclosures coexist with accurate review-status answers, and two answers use a planted false count. Interpretation, notification, and authorized task performance are distinct outcomes.

0.51

SoK: Failure Modes in Common Criteria Product Evaluation - A Taxonomy and Design-for-Evaluability Guidance

cs.CR cs.SE ·2026-10-07 ·Punit Suketu Patel
publishedsurvey
Abstract · 25 pages, 1 figure, 1 table. Accepted at the Security Standardisation Research (SSR) Conference 2026; to appear in Springer LNCS. Author's s…

The Common Criteria (CC; ISO/IEC 15408) is the principal international framework for evaluating the security of IT products, and its certificates gate procurement across government, defense, and regulated industry. Yet evaluations routinely stall, fail, or yield certificates whose assurances do not survive real-world deployment. Prior work has approached this from two directions: economic critiques of misaligned incentives and "security theatre"; and recent data-driven studies that quantify vulnerabilities in already-certified products. Comparatively little systematizes the problem from the evaluator's operational vantage - recurring, cross-vendor failures against the Common Criteria work units themselves. This paper presents a systematization of knowledge (SoK) of failure modes in Common Criteria product evaluation. Drawing on the standard, the Common Evaluation Methodology (ISO/IEC 18045), public NIAP Protection Profiles, and published Security Targets and Certification Reports, we organize recurring failures into a lifecycle taxonomy spanning Security Target scoping, Protection Profile conformance, assurance evidence, cryptographic requirements, functional testing, vulnerability analysis, operational guidance, and post-certification configuration drift. We analyze their root causes, derive a design-for-evaluability framework that product teams can apply before evaluation begins, and identify open problems for cloud, continuous-delivery, and AI-enabled systems.

0.50

Adversarial Images Hijack Web Agents from Visual Grounding to Browser Execution

cs.CR cs.AIcs.CL ·2026-10-07 ·Wanjing Han, Levi Taiji Li, Mu Zhang, Yue Jiang, Guanhong Tao
benchmarkcode/data
Abstract · 20 pages, 8 figures, 6 tables. Code: https://github.com/MoonTea0416/WebMirage

Modern web agents built on large vision-language models process webpages, select relevant UI elements, and translate model outputs into browser actions. Existing visual red-teaming approaches use adversarial visual content to manipulate this process. However, they primarily target model inference and do not explicitly account for structured input processing or action post-processing. Consequently, model-level success does not establish control over browser execution and cannot reliably characterize end-to-end agent robustness. To address this gap, we formulate red teaming for vision-grounded web agents as an end-to-end grounding-to-execution problem, and introduce WebMirage, a framework that crafts localized visual perturbations that cause agents to select attacker-controlled content and execute the corresponding browser action across varying webpage renderings. It uses a role-slot abstraction and webpage recomposition to capture competition among webpage elements, and dataflow analysis to align optimization with action post-processing. We evaluate WebMirage across four agent configurations and six VLM backbones on 2,250 tasks covering 13 public websites and a sandbox benchmark. WebMirage achieves an average attack success rate of 91.9%, compared with 17.4% for the strongest baseline, and remains effective against three agent-level defenses.

0.46

PyCache Trap: The Inspection-Execution Gap in Agent Skill Scanners

cs.CR cs.SE ·2026-10-07 ·Jie Liao, Simeng Qin, Wenqi Ren, Wei Zhou, Junhao Wen +3 more
code/data
Abstract · 28 pages, 5 figures. Code: https://github.com/leo0481/PyCacheTrap

Agent skills combine instructions with executable resources, giving third-party packages access to an agent's runtime. Existing skill scanners inspect documentation and visible source, but Python may execute a bundled bytecode cache with different behavior. We study this gap between inspection and execution through PyCache Trap, which pairs benign source with a substituted cache accepted by the loader and connects it to a task-relevant invocation. Scanner-guided rewriting changes the invocation wording while preserving the cache body, separating package admission from recognition of the concealed behavior. Across 100 skills and seven scanners, PyCache Trap achieves 94-100% attack success, with no semantic recognition of the cache-resident behavior. We propose execution-aware validation (EAV) to connect inspected instructions, scripts, imports, and runtime artifacts in a typed execution graph. EAV combines grounded behavioral analysis with trusted reproduction of compiled artifacts. It detects all 100 evaluated source-present cache substitutions and reaches 92.8% Recall at 10.0% FPR across five attack families and 200 benign skills. The results support checking the executable artifacts a runtime can select as part of skill admission, within the supported loaders and code-object normalization. The code is released at https://github.com/leo0481/PyCacheTrap.

0.43

Detection-Guided Adaptive Purification with Diffusion Models for Robust Audio Deepfake Detection

cs.CR cs.SD ·2026-10-07 ·Muhammed Salih Kayhan, Qiben Yan
top venue
Abstract · Accepted at the 4th EAI International Conference on Security and Privacy in Cyber-Physical Systems and Smart Vehicles (EAI SmartSP 2026)

Audio deepfake detectors remain vulnerable to adversarial perturbations that suppress the acoustic cues used for detection, allowing manipulated utterances to evade the detector. Although existing defenses can improve robustness, they require retraining the detector or introduce additional distortion. Diffusion-based purification instead leaves the pretrained detector unchanged, but existing methods use the same purification strength for all inputs, creating a trade-off between removing adversarial perturbations and preserving the subtle spoofing cues needed for detection. In this paper, we propose Detection-Guided Adaptive Purification (DGAP), a diffusion-based defense that adjusts purification strength per input. Building on the observation that a light purification perturbs the detector score of an adversarial input far more than that of a benign one, the framework uses the resulting score shift as a reference-free indicator of adversarial manipulation. Inputs with small shifts are passed unchanged, whereas flagged inputs undergo stronger purification before final detection. We evaluate the framework against three adversarial attack settings across three deepfake detectors, and compare it with nine existing defenses. Our results show that DGAP achieves the best defense performance across all detectors while leaving benign inputs nearly unaffected, and remains effective under the defense-aware adaptive attack.

0.35

Defensive Sufficiency in a Stackelberg Model of AI Security

cs.CR cs.AI ·2026-10-07 ·Subhabrata Majumdar, Rajlakshmi Chavan
Abstract · 27 pages, 3 figures, 1 table

Feedback from automated testing, human red teaming, and incident response can strengthen an AI system's defenses when discovered failures lead to effective repairs. We study when this feedback process provides sufficient protection and when investing in it is economically worthwhile. We begin by showing that an attack surface composed of finite number of inputs is defended with probability 1 if every unresolved attack has a persistent chance of discovery, repairs are effective, and subsequent updates preserve earlier protection. We derive completion-time bounds and extend the analysis to growing attack surfaces, repairs that generalize across related attacks, and multiple discovery mechanisms. These results distinguish eventual protection against each fixed attack from complete protection at a single time. We then formulate a defender-led Stackelberg game in which the defender invests in proactive discovery and reactive repair, anticipating the attacker's choice of search effort. We characterize the least-cost allocation that deters attack and the equilibrium regimes in which the defender funds neither capability, one capability, or both. Numerical experiments illustrate these regimes and show how faster repair can reduce compromise duration without reducing compromise probability.

0.35

Secure-CUA: Controlling Untrusted Influence in Computer-Use Agents

cs.CR cs.AI ·2026-10-07 ·Sarthak Choudhary, Mihai Christodorescu, Ashish Hooda, Somesh Jha, Tongxin Li +1 more
Abstract

Computer-use agents (CUAs) perform tasks across applications (such as desktops, mobile apps, and web browsers) by observing graphical interfaces and issuing commands such as clicks and keystrokes. These interfaces combine trusted controls and content with untrusted content needed for legitimate tasks. An adversary controlling this untrusted content can embed instructions or misleading visual cues to change the agent's intended action or redirect its commands to the wrong interface target. We formalize security requirements for both the agent's decisions and their execution through GUI commands. In an ideal execution model, we show that enforcing both requirements at each step protects execution traces. We instantiate this model in Secure-CUA, our system for secure CUA execution. Its key idea is to commit to an explicit per-action program, called an $\textit{action transaction}$, before accessing untrusted content. Each transaction fixes its queries to untrusted content and the permitted uses of their responses. The system masks untrusted regions and evaluates each transaction to produce the next action, using an isolated query model to answer its queries. It then locates the intended interface target using the masked interface. Under the model's assumptions, Secure-CUA is secure by design, while generating a new transaction at each step helps maintain high task utility by adapting to changing interfaces. We evaluate Secure-CUA under benign conditions on 400 WebArena tasks using three frontier models across $5$ seeds, yielding $6,000$ execution traces. Secure-CUA achieves an average task success rate of $53.55\%$, compared with $55.12\%$ for Vanilla-CUA and $13.17\%$ for CaMeL-CUA.

0.35

Power Side-Channel Membership Inference Attack on Embedded Machine Learning

cs.CR cs.LG ·2026-10-07 ·Sahan Sanjaya, Prabhat Mishra
Abstract

Membership inference attacks (MIAs) threaten the privacy of machine learning (ML) training data by determining whether a sample was used to train a target model. Existing MIAs rely on model outputs, ranging from prediction probabilities to predicted labels, an assumption that can be restrictive for on-device ML systems with limited or inaccessible outputs. However, suppressing model outputs does not eliminate the data-dependent computations that produce them, which may remain observable through physical side channels. We present PSCMIA, a power side-channel membership inference attack against embedded ML models that can infer membership directly from power traces without requiring prediction probabilities or even the predicted labels. We evaluate PSCMIA across multiple datasets (MNIST, FMNIST, CIFAR10, CINIC10), fully connected (FC) and convolutional neural network (CNN) architectures, and two embedded platforms (STM32F3, XMEGA). PSCMIA achieves ROC-AUC values of up to 0.907 on FC models. For CNN models, the ROC-AUC gap between PSCMIA and probability vector-based shadow MIA ranges from 0.006 to 0.116. Across the FC and CNN evaluations, PSCMIA outperforms label-only MIA in 11 of 16 model-dataset-hardware configurations, demonstrating that physical execution can expose membership information even when conventional model outputs are unavailable through unintended power side-channel leakage.

0.32

On the Reliability of LLM-Based Vulnerability Patching Benchmarks

cs.CR cs.SE ·2026-10-07 ·Dang K Le, Wenxuan Shi, Xinyu Xing
benchmarkcode/datahot direction
Abstract

Large language models (LLMs) have shown strong potential for automated vulnerability patching, but current benchmarks can substantially distort reported performance. Drawing on extensive experience developing, running, and stress-testing such frameworks, we identify under-examined pitfalls across three dimensions: (1) agent-level factors, where prompting, tool availability, and detailed instructions can raise success rates without improving developer-aligned patch quality; (2) framework-level factors, where permission errors, infrastructure bugs, and timeout handling can silently suppress or inflate performance; and (3) dataset-level factors, where bug reports and single proof-of-concept (PoC) tests fail to capture whether patches address root causes or follow developer intent. We curate 112 historical bugs from 84 open-source C/C++, Go, and Rust projects, each with PoC tests, regression tests, and additional developer tests that assess alignment with the original developers' design principles. Through controlled experiments and case studies, we show that LLMs can achieve high PoC passing rates under ideal conditions, yet benchmark execution choices can materially change measured success. More importantly, developer-test passing rates remain low and improve only marginally with newer models, suggesting that models increasingly suppress symptoms without consistently producing upstream-quality fixes. These results show that benchmark scores are highly sensitive to evaluation design, and we provide practical guidelines for more rigorous, reliable, and reproducible evaluation.

0.31

The Hint Weight of ML-DSA Signatures Is Key-Dependent: An Empirical Study across the Three FIPS 204 Parameter Sets

cs.CR ·2026-10-07 ·Dominik Blain
unusual
Abstract · 9 pages, 4 tables

Every ML-DSA (FIPS 204) signature carries a public hint vector h. We find that the Hamming weight of each hint polynomial h_k depends on the signing key: to first order it measures the Euclidean norm of (t0)_k, the low-order part of t that key generation leaves out of the public key. A closed-form model predicts the per-key mean weight with Pearson r between 0.95 and 0.98; once the norm is accounted for, we detect no key-dependent signal above sampling noise. We measure the effect on the reference C implementation, with 200 keys and 2,000 signatures per key for each parameter set. A one-way ANOVA rejects key-independence of the total hint weight for ML-DSA-44, ML-DSA-65 and ML-DSA-87 (F = 30.1, 10.3, 11.1). The effect is weak: the key explains 0.5% to 1.5% of the variance of the total weight. A single signature identifies its key among 200 with 1.1 to 1.3 times chance accuracy from the total weight, and 1.5 to 1.8 times from the per-polynomial weight vector. A one-sided test at significance 0.001 separates two typical keys with probability one half after about 1,300 to 3,700 signatures per key. The hint weight does not endanger the signing key. The Dilithium designers do not treat t0 as secret, and t0 is known to be recoverable from signatures. The hint weight is, however, a weak statistical fingerprint: with enough signatures, it can be used to test whether they come from a common key, without the public key. Bounded-weight signing, a backward-compatible filter whose effect on the security argument we did not analyze, removes 86-91% of the between-key spread of the total weight but leaves the per-polynomial channel largely intact.

0.30

MRCert: Towards Post-deployment Patch Robustness Certification for Adversarially Patched Samples via Type-specific Masking

cs.CR cs.AIcs.LGcs.SE ·2026-10-07 ·Qilin Zhou, Zhengyuan Wei, Haipeng Wang, Zhuo Wang, Shuo Liu +1 more
Abstract

In post-deployment time, inputs to deep learning models may or may not be adversarially patched. Patch robustness certification on such inputs within a patch bound can verify their label benignity and should retain high prediction accuracy. However, existing smoothing-based and masking-based recovery defenders cannot achieve both simultaneously: they degrade the prediction accuracy much and cannot verify the benignity of the returned label of an adversarially patched input, respectively. We propose MRCert, the first masking-based certified recovery defender that shows the feasibility of achieving both. Unlike all existing works to apply a common condition across both types of input (benign and adversarially patched samples) for certification, MRCert infers type-specific necessary properties of deep learning models for both types in post-deployment time and formally relates them to verify the label benignity through a novel type-oriented design of label recovery and certification function pair. Without incurring the degradation in clean accuracy caused by smoothing, experimental results confirm that MRCert achieves 35.1\% adversarial certified accuracy on ImageNet at patch size 16 pixels, whereas the SOTA PatchCURE fails completely.

0.30

Ask the Expert: LLM-Guided Reinforcement Learning for Autonomous Cyber Defense

cs.CR cs.LG ·2026-10-07 ·Fernando Martinez, Abhishek Satyam, Tao Li, Junaid Farooq, Ying Wang +1 more
published
Abstract · Accepted for publication at IEEE GLOBECOM 2026. Proceedings forthcoming. 6 pages, 4 figures

Policy-based reinforcement learning (RL) approaches have produced promising results for autonomous cyber defense; however, they are sample-inefficient in settings where defenders must respond under delayed, partial observations with actions from large action spaces. While large language models (LLMs) may reason semantically about security state space, high latency and trust assumptions prevent attractive in-line deployment models. We introduce Ask the Expert, a training-time guidance framework which first summarizes hard cyber-defense states, then intermittently queries an LLM for host-level defensive recommendations via a constrained action interface, and finally transforms those recommendations into tiered reward shaping for use with PPO. Because the LLM is discarded after training, deployment is a pure RL policy. Across TTCP CAGE CC1 and CC2 and both attacker types, this asymmetric design improves sample efficiency over PPO and outperforms the evaluated potential-based reward shaping (PBRS) baselines, while retaining the strongest terminal mean and requiring no LLM dependency at deployment time.

0.30

DITTO: A Context-aware Pickle-based Pre-Trained Model Scanner for Effective Security Audits

cs.CR cs.SE ·2026-10-07 ·Qiaolin Qin, Wanpeng Li, Benoit Baudry, Lorenzo De Carli, Heng Li +1 more
benchmark
Abstract

Pre-trained models (PTMs) are widely distributed as serialized binaries, but their reuse often exposes software supply chains to deserialization attacks. Despite the emergence of safer serialization formats, the unsafe Pickle format remains prevalent: our analysis of over 10,000 popular Hugging Face repositories reveals that 9.3% rely on Pickle. While many defense mechanisms have been proposed, state-of-the-art model scanners suffer from a coverage-precision gap, missing security-sensitive behaviors and generating excessive false alerts. In this paper, we introduce DITTO, the first stack-based, context-aware scanner for Pickle-based PTMs. DITTO faithfully tracks Pickle virtual machine state transitions and performs context-aware semantic analysis to infer model intentions. We also present PickleBench, a benchmark of 959 benign and 92 malicious real-world models, including extension registry attacks previously missed by existing tools. Across multiple evaluations, DITTO achieves 100% scanning coverage, a 0% false-negative rate, and a 0.7% false-positive rate, yielding an F1 score of 0.966, significantly outperforming state-of-the-art scanners. By minimizing false alerts while preserving detection accuracy, DITTO generates actionable security reports with contextual evidence, enabling safe PTM reuse and strengthening software supply chain integrity.

0.29

Package Hallucination Attacks on Coding Agents through Prompt Injection in Rule Files

cs.CR cs.AI ·2026-10-07 ·Yupu Wang, Zhengyuan Jiang, Reachal Wang, Neil Zhenqiang Gong
Abstract

Modern agentic coding frameworks increasingly rely on community-shared rule files (e.g., AGENTS.md or .cursorrules) to guide autonomous code generation, yet the security risks of this pipeline remain underexplored. To bridge this gap, we introduce the package hallucination attack, where an attacker injects malicious prompts into benign rule files to induce coding agents to replace legitimate dependencies with attacker-controlled packages. To obtain effective malicious prompts injected into rule files, we propose PackHallu, an evolutionary optimization framework that iteratively rewrites these injected prompts using trajectory-level feedback and LLM-guided mutations. Evaluations across multiple benchmarks, LLMs, and agent frameworks show that PackHallu achieves high attack success rates and strong transferability across diverse models and agent combinations. Our findings demonstrate that coding agents are vulnerable to package hallucination attacks, highlighting the urgent need for stronger security safeguards in autonomous coding systems.

0.28

Trust a Few: The Weakest Assumptions a Protocol Needs

cs.CR ·2026-10-07 ·Bhumika Mittal, Aalok Thakkar
unusual
Abstract

Protocol verifiers check whether a protocol meets a security goal under stated trust assumptions, such as that a key is never leaked, a value is fresh, or a channel is authentic. They do not say which of those assumptions the goal needs. Rowe, Guttman and Liskov asked for the weakest assumptions under which a protocol achieves a goal and left the question open. We answer it for assumptions about keys, values and channels. Call a run that violates the goal an attack, and the assumptions that would rule it out its stopping set. The least a protocol must trust to meet a goal is exactly the set of minimal ways to stop all of its minimal attacks. Several such sets may exist. The answer becomes unique once either/or assumptions are allowed, and it is a single set exactly when every minimal attack is stopped by a single assumption. A Galois connection between assumptions and goals explains why: a goal may end in "or", but its hypothesis may not. The same structure gives the trust needed by a conjunction of goals and an exact condition under which composed protocols need no extra trust. To compute the answer, a loop asks a verifier whether a candidate suffices, records what stops each attack it reports, and recomputes the candidates. Deciding whether some trust of at most k assumptions suffices is NP-hard. With a bounded analyser of our own and with CPSA, the loop finds the weakest trust for 18 goals over ten protocols and their variants, and on each it agrees with evaluating every trust using the same verifier. The answers include an assumption that our model of the adopted fix of Kerberos PKINIT states but the client's authentication of the server does not need, and channel assumptions under which Needham-Schroeder meets its goal in a bounded model.

0.27

Faster PMNS Multi-precision Multiplications Using Truncated Montgomery Technique

cs.CR ·2026-10-07 ·Laurent-Stéphane Didier, Alexy Dutois-Ruiz, Jean-Marc Robert
Abstract · 21 pages, one figure

The Polynomial Modular Number Systems (PMNS) aim to represent elements of fields or rings of large characteristics using polynomials satisfying bounds on some parameters (degree, absolute values of the coefficients). Those PMNS, while some conditions are fulfilled, using convenient parameters and implementation features, allow some speed-ups in cryptographic computations. Recent works (Meloni \emph{et al.} \cite{MeloniPV25}) propose better speed-ups by improving the parameter generation of the system, and by using multi-precision polynomial coefficients, i.e. coefficients stored using several machine words. In this work, we first present PMNS schoolbook software implementation better in some context than the Toeplitz state-of-the-art counterpart of \cite{MeloniPV25}, taking advantage of smaller memory cost by being free to choose a smaller $n$ parameter, minimizing at the same time the complexity of the multiplication. This implementation for 4096 bit modulo size shows element size smaller by 11\% and 20 \% speed-up. We thus present a new improvement on the PMNS, applying to the \textsl{internal reduction} an approach similar to the truncated Montgomery reduction technique presented by Didier \emph{et al.} in \cite{DidierEGR24}. In the context of software implementations using \texttt{AVX512} instruction set extension, and modulo size up to 8192 bits, this new approach allows speed-ups up to 15\% in modular multiplication computation, in comparison with conventional approaches.

0.26

CPU-Auth: Device Fingerprinting for Authentication via DVFS Side-Channel

cs.CR cs.LG ·2026-10-07 ·Ryan Swift
Abstract · 13 pages

Lack of effective authentication has resulted in numerous security and privacy breaches, including unauthorized access to protected information, identity theft, and fraud. One approach to mitigating such attacks is Multi-Factor Authentication (MFA), in which users must provide multiple pieces of information for authentication. Some secondary authentication factors include SMS text verification codes, biometrics, and tokens. Each contains at least one notable flaw: SMS is notoriously insecure; biometrics rely upon access to sensitive personal data; and tokens require dependence on third-party providers (e.g. OAuth providers). This work explores CPU-Auth, a novel authentication mechanism based on unique variations in the physical characteristics of the CPU of a computing device. By measuring the behavior of the Dynamic Voltage and Frequency Scaling (DVFS) governor remotely from within a browser, unique properties of the CPU can be leveraged to establish a hardware-based device fingerprint for use in CPU-Auth. The performance of CPU-Auth is evaluated on over 50,000 data traces using distance-based and deep learning methods. CPU-Auth is part of a larger research project, and the results provided in this report reflect only the contributions made to the project by members of this group.

0.26

When Flaws Cascade: Understanding Vulnerabilities and Exploitation Chains in JavaScript Engines

cs.CR cs.SE ·2026-10-07 ·Yuhan Ma, Jiongchi Yu, Xiaofei Xie, Qiang Hu, Zhiyi Zhang +1 more
datasethot direction
Abstract · 10 pages

JavaScript engines are pivotal to modern web browsers, enabling the execution of dynamic and interactive web applications. However, their complexity and widespread adoption make them prime targets for attackers exploiting vulnerabilities. While existing research has focused on detecting vulnerabilities of JavaScript engines, a significant gap remains in systematically understanding the characteristics of these vulnerabilities, including their symptoms, root causes, and exploitability. This paper bridges this gap by presenting the first comprehensive empirical study on vulnerabilities in JavaScript engines, investigating their characteristics and potential exploitation strategies. We construct a dataset comprising 241 vulnerabilities across four mainstream JavaScript engines from 2017 to 2024. Through in-depth analysis, we first develop taxonomies for symptoms and root causes. Building on this understanding, we investigate the exploitability of these vulnerabilities, identifying key prerequisites and extracting vulnerability trigger chains that demonstrate how logical errors propagate into memory safety violations. Additionally, we analyze the mitigation strategies to counter these exploits. Finally, we summarize key implications for various stakeholders, including developers and researchers, offering actionable insights to improve the security and resilience of JavaScript engines.

0.26

Defining Purpose-Limited Secrets

cs.CR ·2026-10-07 ·Bhumika Mittal, Aalok Thakkar
Abstract

A cryptographic secret is issued for a purpose but grants a capability, and the capability is usually larger. Deployed systems state the purpose in policy and enforce only the capability. We make the purpose a property of the secret. A mechanism sees operations, not reasons, so a scheme enforces an admissible set $A(P)$ of operations that stands in for a declared intended use $I$; whether $I$ captures the human purpose is a modeling obligation that marks where policy takes over. Against the same $I$ we define three regimes: under confinement an operation outside the intended use is infeasible, under mediation a trusted component refuses it, and under accountability use beyond a stated bound is possible but attributed to the holder. The confinement games, for unpredictability and indistinguishability, coincide with the key-query forms of functional-signature unforgeability and constrained-PRF pseudorandomness and correspond to simulation-secure functional encryption within its feasibility boundary; stated against $I$, they also register a predicate that permits too much. A payment token uses all three regimes on one secret: it pays one capped charge at one merchant, can be narrowed but not widened, is checked by the merchant, and identifies the withdrawing account if spent twice. Attenuation unforgeability reduces to unforgeable signatures and a collision-resistant hash without random oracles; traceability and non-frameability hold under discrete log and unforgeable signatures in the random-oracle model. We present this framework as a common specification against which existing primitives are stated and compared.

0.25

Receiver-Domain Behavioral Probing for Backdoor-Resilient Federated GPS Spoofing Detection in UAV Networks

cs.CR eess.SP ·2026-10-07 ·Will Jedrzejczak, Cole Walther, Dilpreet Gill, Moh Khalid Hasan
Abstract · This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may…

Federated learning (FL) allows unmanned aerial vehicle (UAV) fleets to train a shared GPS spoofing detector without exchanging raw receiver observations. Some UAV-FL schemes use client-reported performance to determine aggregation weights, which allows compromised clients to manipulate their influence on the global model. We study an attack in which two of ten clients poison part of their data, scale their model updates, and inflate their reported accuracy. The accuracy-weighted rule produces a backdoor lift of +0.3036, compared with +0.2415 for attacked FedAvg. To address this vulnerability, we develop receiver-domain behavioral probing. The coordinator evaluates submitted models on counterfactual spoofed samples constructed by assigning benign values to discriminative GPS features and computes trust from the models' responses without using reported accuracy. With independent and identically distributed (IID) client data, our method reduces backdoor lift to -0.0265 and flags the compromised clients. It assumes an honest majority but does not require a configured Byzantine-tolerance parameter. When the number of attackers exceeds the tolerance configured for Multi-Krum, its lift increases from +0.0061 to +0.2837, whereas our method remains close to the honest baseline. We evaluate the approach using a public single-receiver dataset partitioned into simulated UAV clients and examine its limitations under strong client heterogeneity.

0.25

Automotive Hardware Attacks: An Architect's Guide to TARA

cs.CR ·2026-10-07 ·Jakub Breier, Xiaolu Hou
Abstract

Threat analysis and risk assessment (TARA) according to ISO/SAE 21434 treats implementation-level hardware attacks inconsistently. We show that two of the three attack-feasibility approaches exemplified in the standard rate every attack path that requires physical access as very low by construction, independent of the implementation, while the third can assign the same path any of the four ratings, depending on the assumed state of attacker knowledge and on the attacker profile. This paper presents a hardware-aware extension of the TARA that covers side-channel analysis (SCA), fault-injection attacks (FIA), and abuse of debug and test interfaces. It adds an attacker model for physical access over the vehicle lifecycle; a path-dominance relation on the attack-potential factors that identifies when a simpler path makes a sophisticated one irrelevant; a three-level hardware relevance gate (H0 to H2) with an explicit decision rule that fixes whether resistance of a hardware step may be argued, must be documented, or must be demonstrated by test; and a rating rule under which a hardware step receives no credit for implementation-specific resistance until the required evidence exists. The gate is positioned relative to cybersecurity assurance levels and targeted attack feasibility. The method is applied to a reference gateway architecture, and its structural inputs are examined on ten published attacks on automotive components. In seven of the ten studies, the demonstrated attack included a debug, programming, boot, diagnostic, or update mechanism, whereas only four studies included an attack on a cryptographic computation.

0.24

Safe at One Loop, Risky at Another: Aligning Safety Across Recurrent Depths in Looped Language Models

cs.CR cs.LG ·2026-10-07 ·Yi Wang, Xiuyuan Qi, Dongqi Han, Dongsheng Li, Wenjie Wang
Abstract · 25 pages, 6 figures

Looped Language Models (LoopLMs) provide a parameter efficient approach to scaling model capabilities through repeated use of shared parameters across recurrent steps. Since each recurrent depth can be read out independently, a single LoopLM exposes a broader output space across inference depths, raising an important question: whether safety is preserved throughout recurrent computation. Prior evaluations suggest that deeper recurrence can improve safety on harmful queries, but robustness under jailbreak attacks remains unclear. We therefore conduct a comprehensive safety evaluation of LoopLMs under jailbreak attacks targeting different recurrent depths. We find that attack success can increase at deeper inference depths, the same query can elicit different safety behaviors across depths, and attacks constructed against one depth can transfer to others. Moreover, SFT and preference alignment do not eliminate these safety gaps, motivating an alignment method designed for LoopLMs. We introduce SafeBridge, which combines lightweight depth specific control of shared recurrent layers, selective state bridging, and joint safety supervision across recurrent depths. Across model scales, multiple attack methods, and safety benchmarks, SafeBridge substantially reduces attack success for both matched-depth and cross-depth attacks. It also improves general utility over the vanilla models while maintaining comparable over-refusal behavior. Our results show that the safety of a LoopLM cannot be inferred from a single recurrent depth, motivating safety alignment across recurrent computation. Our code and model checkpoints will be released upon acceptance.

0.23

Constitution-Guided Watermarking

cs.CR cs.CLcs.LG ·2026-10-07 ·Toluwani Aremu, Samuele Poppi, Nils Lukas
Abstract · Working paper (under review)

Regulation such as the EU AI Act requires providers of language models to mark generated text in a machine-readable way. Watermarking meets this requirement by embedding a secret-key signal during generation, so the provider can later detect its model's text. A watermark should stay detectable, preserve quality and survive paraphrasing, but a stronger signal survives more paraphrasing and distorts more text. Providers resolve this trade-off with one configuration for all requests, although a provider may want robustness for misuse-prone requests and no watermark for editing assistance, which the EU AI Act exempts. We introduce \emph{constitution-guided watermarking}, in which the provider states these requirements as natural-language rules, a \emph{watermark constitution}, and each request receives its rule's configuration. Offline, a reasoning agent reads the rules and the watermark code and refines each rule's configuration from measured detection, quality and robustness. At deployment, a monitor routes each prompt to its rule's configuration without modifying the serving model. Each configuration is stored with its evidence, so the provider can audit every deployment decision, including unmet requirements. Our evaluation with the KGW watermark shows that (i) on misuse-prone requests, the framework detects more paraphrased text than fixed configurations and random search at matching clean detection and quality, (ii) it exempts all editing requests, and (iii) changing the stated priorities shifts the selected configuration as intended.

0.23

BetweenCut: Private Heavy-Node Classification with Doubly Logarithmic Error in Tree Height

cs.CR cs.DS ·2026-10-07 ·Ergute Bao, Graham Cormode, Xiaokui Xiao, Ting Yu
Abstract

Finding heavy nodes in a tree---those whose counts exceed a given threshold---is a building block for analysis and learning over structured data. Achieving record-level differential privacy (DP) without sacrificing accuracy is challenging because each record contributes to counts along an entire root-to-leaf path, allowing privacy costs to accumulate across levels. Existing methods account for the multiple threshold comparisons for each record incur additive error margins of $Ω_{\varepsilon,δ}(\log h)$ or $Ω_{\varepsilon,δ}(\sqrt{\log h})$ for tree height $h$. We introduce \textsc{BetweenCut}, an $(\varepsilon,δ)$-DP algorithm with an additive error margin of $O_{\varepsilon,δ}(\log\log h)$, improving the existing bounds for deep trees. This error holds simultaneously for all nodes and is independent of the input database size.

0.23

Black-Box Adversarial Patch Attacks on VLAs via Ancestor VLM Exploitation

cs.CR ·2026-10-07 ·Xiaoyi Pang, Haoyue Feng, Quanxin Shou, Yikun Miao, Zhengyang Yan +1 more
Abstract

Vision-Language-Action models (VLAs) are increasingly deployed in safety-critical physical environments, yet their adversarial robustness remains poorly understood. Existing attacks typically assume white-box access or rely on surrogate VLAs, which rarely holds in real-world deployments. Our key insight is that most VLAs are adapted from a publicly released pretrained vision-language model (VLM), inheriting two capabilities essential for action generation: visual perception and instruction-conditioned grounding. Therefore, this paper explores a previously unaddressed question: can an adversary attack deployed VLAs using only their ancestor VLMs? To this end, we propose three adversarial patch attacks that disrupt the inherited capabilities: a vision disruption attack that corrupts the projected visual tokens through relative and absolute terms, an instruction-grounded semantic evidence suppression attack that removes the visual evidence required for instruction-grounded concepts, and a joint attack that unifies both objectives under a two-phase curriculum. Experiments across different VLA families on both simulation and static real-world images show that patches optimized on the ancestor VLM cause substantial degradations in VLA task success rates, demonstrating that VLAs inherit adversarial vulnerabilities alongside their foundational capabilities. This effect is not uniform: it is strongest on tasks that require precise instruction-grounded localization, and nearly vanishes on policies whose adaptation rewrites the shared visual-semantic representation or whose action head iteratively smooths perturbations away. By characterizing the boundary conditions of vulnerability inheritance and providing analysis of why the inheritance effect holds or fails, we advance the understanding of safety for VLA-involved systems.

0.23

MARS: Malware Analysis with Rule-Based Scoring of LLM Claims

cs.CR cs.AI ·2026-10-07 ·Hyeongjun Choi
Abstract · 17 pages, 3 figures, 12 tables

Large language models can triage malware through direct verdicts or behavioral claims scored by an external policy. We present MARS, a malware triage framework, and compare direct classification with single-pass claim scoring using the same evidence collector and identical static evidence bundles for each model. The evaluation covers 1,195 PE and ELF binaries grouped into 1,001 near-duplicate clusters and six language models, with deterministic rules providing a baseline. Direct classification is more accurate for all six models. On samples with usable outputs from both paths, its accuracy advantage ranges from 3.7 to 20.9 percentage points, with all 95% cluster-bootstrap confidence intervals for the differences above zero. It also achieves higher malicious alert recall in ten of twelve platform and model combinations. Claim mediation provides no consistent reduction in performance variation across models. Separate subset studies find more consistent alert decisions for direct classification and a larger recall loss for the claim path when predefined indicator fields are removed. In a family identification probe, claims yield higher accuracy than verdict labels but lower accuracy than evidence text. Retained claims expose the inputs to verdict computation and permit policy revision without another model call. We reproduce archived verdicts exactly and apply a revised policy to the same records, including outputs from two additional models withdrawn by their provider. Under the evaluated claim taxonomy and additive policy, these results favor direct classification when only a verdict is required, while demonstrating that retained claims support explicit policy inspection and revision.

0.22

BRANCH: Bypassing Multi-Scanner AI Guardrails

cs.CR cs.AI ·2026-10-07 ·William Hackett, Peter Garraghan
Abstract · 13 pages, 21 figures, 4 tables

AI systems increasingly rely on Large Language Models (LLMs) as core reasoning engines, making them targets for prompt injection and jailbreaks. Guardrails monitor and validate model inputs and outputs, yet their isolated, task-focused detection leaves gaps in their classification making them susceptible to bypasses. In response, guardrail systems formed by multiple scanners have emerged that collaboratively detect different types of malicious instructions, whereby shared latent representations across classification boundaries render established bypassing techniques ineffective. We propose BRANCH, a bypassing methodology designed for multi-scanner guardrail systems. Our method leverages a branching tree search approach that dynamically applies adversarial perturbation against individual scanners, with subsequent perturbation optimization and technique selection based on overall improvement across all guardrail system scanners, effectively decoupling bypass evaluation from attack signal optimization. Our findings demonstrate that BRANCH achieves 100% attack success rate across 6 guardrail systems in 120 scenarios with 72% fewer queries and 4.5x reduced wallclock time compared to established techniques, while preserving semantic meaning within the bypass. We also show how bypasses generated by BRANCH transfer to 29 unseen guardrails, including 8 commercial black-box guardrails, improving attack success in some cases up to 100% with no additional optimization.

0.21

From Investigation Failures to Reliable SOC Agents: Understanding and Improving LLM-Based Alert Triage

cs.CR cs.AIcs.MA ·2026-10-07 ·Saimon Amanuel Tsegai, Alex Kantchelian, Danfeng, Yao, Peng Gao
benchmark
Abstract · Preprint

Security operations centers (SOCs) must triage large volumes of alerts, most of which are benign, while missed attacks can remain uninvestigated. Tool-using large language model (LLM) agents can retrieve evidence during triage, but it remains unclear how reasoning strategies determine what to gather and when an investigation is sufficient to close an alert. We study five representative approaches spanning single-pass tool use, iterative retrieval, sampled investigations, self-review, and explicit verification. To support this study, we build ALERT-BENCH, an interactive benchmark that replays enterprise telemetry through a live SIEM and requires each system to retrieve evidence. Across 1,247 alerts from a multi-stage attack scenario, every approach missed at least 40.4% of attack-related alerts. Trace analysis shows that attack alerts are more likely to be dismissed when searches return no records, same-context review has negative net correction, and dismissal receives no consistently stronger investigation than escalation. Based on these findings, we further design AIDA (Adversarial Investigation and Dialectical Analysis), a multi-agent framework that requires an explicit proposed decision before independent challenge and stronger evidentiary requirements before dismissal. AIDA preserves investigation history in an append-only Investigation Ledger and keeps the challenge in a separate reasoning context. A separate Judge adjudicates the proposed decision and challenge against evidence, resolving the alert or requesting another round when evidence is missing. On the same alerts, AIDA achieves an F1 score of 0.958, compared with 0.371-0.744 for the studied approaches, and reduces the false-negative rate from 40.4% to 3.1% while escalating 18.4% of alerts to analysts. These results show that structuring evidence retrieval and decision review can substantially improve agentic SOC triage.

0.20

Correct Answers, Unsupported Findings: Evidence Binding in Forensic Reconstruction of LLM Agent Logs

cs.CR cs.AI ·2026-10-07 ·Taehyeon Yun, Dongho Kim, Geonwoo Kim, Juyoung Seo, Minseok Hur +1 more
Abstract · 11 pages, 2 figures

Forensic reconstruction of LLM-agent actions requires not only recovering the correct value, but establishing which preserved record supports that finding. Tool logs, generated explanations, and local citation identifiers capture different parts of this evidence, yet a citation identifier does not establish a source unless its binding to a record is preserved. We audit this distinction using 64 mechanically checkable cases from saved AgentDojo Banking executions. Two LLM readers reconstruct source relationships under controlled variations in visible evidence and identifier-to-record bindings. We separately evaluate complete-record agreement, evidence-grounded findings, justified abstention, and unsupported assertions. With original identifiers and no binding table, Sonnet recovered every literal source location but made unsupported citation-source assertions in 26 of 28 cases requiring the missing relation; 22 nevertheless matched the complete reference. Adding explicit bindings improved grounded reconstruction for both readers, whereas identifier renaming alone provided no consistent remedy. A deterministic same-packet comparator correctly resolved the bounded task or abstained throughout. These results show that factual agreement alone is insufficient for evaluating forensic reconstruction of agent logs and motivate preserving explicit record bindings to distinguish supported findings from correct guesses.

0.19

Pump-and-Dump meets Honeypot Tokens: Detection and Analysis of Telegram Bait-and-Trap Schemes

cs.CR cs.CY ·2026-10-07 ·Federico Cernera, Massimo La Morgia, Alessandro Mei, Francesco Sassi
Abstract

While Pump-and-Dump schemes have been extensively studied on centralized exchanges (CEXs), how they operate in decentralized exchanges (DEXs) remains largely unexplored. We monitor 83 Telegram channels used to coordinate Pump-and-Dump campaigns and collect 3,677 events across the BNB Smart Chain and Ethereum. Our analysis reveals that, although these operations superficially resemble CEX-based Pump-and-Dump schemes, their underlying mechanism is fundamentally different. Rather than manipulating prices, organizers orchestrate deceptive Pump-and-Dump campaigns around honeypot tokens whose smart contracts allow Telegram subscribers to buy but prevent them from selling. Unaware of this restriction, victims purchase these tokens and are unable to recover their funds. We refer to this new fraud mechanism as Bait-and-Trap. We show that Bait-and-Trap provides a substantially more reliable profit strategy than Pump-and-Dump schemes: organizers gain in 99.3% of cases, extracting over $7 million in profit. To counter this threat, we develop a transaction-simulation tool that detects honeypot tokens before purchase by testing whether a user can successfully sell against the token's live contract state, providing a practical defense against Bait-and-Trap operations.

0.15

Speedbumps: Rejection Attacks on Speculative Decoding

cs.CR cs.AI ·2026-10-07 ·Adam Y. J. Jones, Yu Yuan, Sergio Maffeis
Abstract

Speculative decoding is a popular technique for increasing the speed and reducing the costs of large language model (LLM) inference by verifying multiple draft tokens in a single target-model forward pass. The resulting benefit depends on the ability of the drafter to approximate the target model's distribution. In this work, we study Speculative Rejection Attacks (SRAs), a novel class of attacks that cause draft and target models to disagree more often, resulting in fewer draft tokens being accepted per draft cycle. This leads to more target model forward passes needed per generated token, slowing down inference and increasing costs for the victim. We introduce two attacks which append an adversarial suffix to attacker-controlled content to degrade speculative decoding on a victim's prompts. Both attacks optimise the expected length of the accepted speculative prefix, estimating per-depth acceptance from the target's probability of the drafted proposals (Speedbump-P) or from the overlap between the draft and target distributions (Speedbump-D). In some cases, attacks degrade speculative decoding to the point of being slower than autoregressive decoding. The degradation reduces the output quality - regularisation restores output quality but gives up most of the degradation, trading effectiveness for stealthiness. Additionally, the suffixes remain effective under sampling, and transfer across drafters (Speedbump-P) or across target models sharing a drafter (Speedbump-D). These findings identify the draft-target interaction of speculative decoding as a realistic attack surface through which adversarial inputs can inflate inference costs.

0.14

Collusion-Secure Semi-Quantum Secret Sharing Scheme using a Quantum Third Party

cs.CR quant-ph ·2026-10-07 ·Santanu Majhi, Nirupam Basak
Abstract

The security of a secret sharing protocol is compromised if one or more dishonest participants can reconstruct the secret by deviating from the prescribed protocol, particularly when such malicious behavior remains undetected. Semi-Quantum Secret Sharing (SQSS) is a variant of secret sharing in which the participants possess only classical capabilities, such as preparing and measuring qubits in the computational ($Z$) basis, while a quantum-capable third party assists the dealer in generating and distributing quantum shares of a classical secret. The limited quantum capabilities of the participants in SQSS protocols may introduce vulnerabilities to collusion attacks, enabling an assisting quantum third party, in collaboration with a dishonest classical participant, to jointly recover the secret even without detection. In this work, we propose a novel SQSS protocol that eliminates this vulnerability and achieves information-theoretic security against collusion attacks. We further prove that the proposed protocol is secure against a broad class of external and internal attacks. Finally, a comparative analysis demonstrates that our construction advances existing SQSS protocols by simultaneously optimizing qubit efficiency, involves a classical dealer, and resilience against DCNA attacks using basic Bell states.

0.12

Pareto-optimal quantum kernel selection for unsupervised anomaly detection on real malware beaconing data

cs.CR quant-phcs.LG ·2026-10-07 ·Boaz Micah, Nadia Milazzo, Maissa Beji, Borja Aizpurua, Llorenç Espinosa-Portalés +6 more
Abstract · 13 pages, 5 figures

Quantum kernel methods are leading candidates for a practical quantum advantage in machine learning, but assessing that potential requires two quantities usually reported separately: how well a kernel performs on the task, and how far its geometry departs from the classical kernels available for the same problem. We introduce a fully unsupervised, multi-objective protocol that optimises simultaneously the normalised pseudo discrepancy (NPD), a label-free proxy for anomaly detection quality, and the geometric difference (GD) to a tuned classical reference kernel, selecting models from the resulting Pareto front. We apply it to malware beaconing detection in real network traffic, using a one-class support vector machine with fidelity and projected quantum kernels over four data encodings, on simulators and on IQM's 20-qubit Garnet processor. NPD-guided selection alone finds a fidelity kernel that beats the tuned classical baseline, but with a geometric difference too small to certify the gain as quantum. Projected kernels reach far larger geometric differences; the Pareto-selected one only marginally exceeds the baseline (AUC $0.782$ versus $0.765$, $g_{C\to Q}\approx 89>\sqrt{N}$ relative to that reference kernel), still below the NPD-selected fidelity kernel ($0.840$).

0.09

Systems & Networking 42

Correspondences as Decisions: JevNexus for Decision-Centric Schema Matching

cs.DB cs.AIcs.LG ·2026-10-07 ·Runze Li, Hanchen Wang, Ying Zhang, Wenjie Zhang
benchmarkcode/data
Abstract · 12 pages, 9 figures

Schema matching increasingly uses generative language models to rerank retrieved column candidates, although the underlying task is a bounded correspondence decision. We present JevNexus, which combines typed pairwise decisions with schema/instance evidence and invokes listwise refinement only when the evidence disagrees and the fused margin is small. The evaluation covers 561 cases from six benchmark families. JevNexus obtains dataset-macro MRR and Hits@1 of 0.930 and 0.909, compared with 0.926 and 0.903 for Magneto, while reducing mean latency from 123.452 to 15.929 seconds (7.750). Paired analysis finds no statistically significant difference in either MRR or Hits@1. The gate invokes listwise refinement for only 5.665% of source columns and avoids the degradation caused by unconditional refinement. Code and experimental artifacts are available at https://github.com/RazeenLI/JevNexus.

0.61

ReSAFT: An Efficient Stuck-at Fault-Tolerant Scheme for ReRAM-based Process-in-Memory Accelerators

cs.AR ·2026-10-07 ·Aniseh Dorostkar, Hamed Farbeh, Hamid R. Zarandi
publishedhighly cited
Abstract

Analog ReRAM-based process-in-memory (PIM) accelerators provide high parallelism and energy efficiency for deep convolutional neural networks (CNNs) inference. However, their susceptibility to permanent faults, such as stuck-at high (SaH) and stuck-at low (SaL) resistance states, poses a major challenge by permanently corrupting the CNN weights mapped to conductance values of ReRAM cells and degrading inference accuracy, which leads to system unreliability in safety-critical applications. In this paper, we propose a fault-tolerant scheme for analog ReRAM-based PIM accelerators to tackle stuck-at faults (SAFs) with minimal redundancy overhead to recover classification accuracy degradation. The proposed scheme contains a redundancy-based hardware solution alongside fault-aware mapping method for ensuring reliable analog computation in ReRAM crossbar. We analyze the impact of varying number of redundant rows and columns on accuracy and design metrics. Subsequently, a multi-objective optimization (MOO) problem is formulated and solved to efficiently determine the number of redundant rows and columns, considering trade-offs among various design metrics. Furthermore, a fault-aware weight mapping is proposed for dual-crossbar structures to further compensate for the accuracy degradation caused by SAFs. Simulation results show that, for the SimpleNet model using the MNIST dataset, the inference accuracy is recovered by approximately 22.39%, on average, across four configurations of optimal solutions, each offering a trade-off between reliability and area, energy consumption, and latency overheads. The mean-time-tofailure (MTTF) improves by about 61x on average compared to the baseline. These selected configurations also reduce energy and area overheads by 32%, on average, in comparison to row-only and column-only configurations.

0.57

The Case for Synthesized Data Compression

cs.DB ·2026-10-07 ·Ivan Iachnyk, Mihail Stoian, Luisa Neubauer, Ryan Marcus, Andreas Kipf
Abstract

Compression is a fundamental tool in data management. Today's compression schemes are generally designed by hand to apply to many different types of data (e.g., run length encoding or arithmetic encoding), leading to a zoo of general-purpose algorithms which work well for many datasets, but may not be optimal for any specific dataset. Inspired by recent advancements in agentic coding, we propose synthesizing custom-tailored data compression algorithms on a per-dataset basis. Unlike prior work using agents to synthesize database components, a data-specific compression algorithm can be fully verified at construction time, alleviating most correctness concerns. Modern file formats, like AnyBlox and F3, allow including the synthesized code alongside the data itself. Experimentally, we show that our synthesis agent can expand the compression-factor/decompression-speed Pareto front, outperforming general-purpose methods by nearly $2\times$ in compression factor.

0.51

From Echoes to Evidence: Cooperative Fusion for Integrated Sensing and Communication

cs.NI ·2026-10-07 ·Xingqin Lin
Abstract · 13 pages, 10 figures, 4 tables

Cooperative integrated sensing and communication (ISAC) can support shared situational awareness for unmanned aerial vehicles (UAVs), road vehicles, and industrial mobile robots such as automated guided vehicles (AGVs). Yet multipath can make nodes agree on a virtual target while useful weak echoes remain unexploited. This paper develops cooperative fusion that separates association confirmation, state refinement, and residual recovery. A cellular reference signal model accommodates arbitrary node counts, three-dimensional motion, monostatic and bistatic links, nuisance parameters, and correlated errors. We characterize report distinguishability under bounded bias and shared-reflector ambiguities that survive additional nodes. Minimax estimation yields directional weak-report refinement that preserves confirmed associations and limits bias-sensitive innovations; acquisition accounts for request and upload costs. Controlled experiments accompany UAV evaluations under a Third Generation Partnership Project (3GPP) stochastic channel model and a ray tracing environment representing San Francisco. In the latter environment, matched-budget fusion raises detection from observable site selection's 88.75% to 92.46% and reduces the false-output fraction from 11.06% to 0.40%. Against the strongest full-band fixed site, the 90th-percentile position error on commonly detected targets falls from 8.54 to 3.50 m. Association controls, local rejection ablations, and refinement studies distinguish the sources of these gains.

0.48

PCDT: A Predictive Cognitive Digital Twin Framework for Intelligent and Autonomous 6G Network Ecosystems

cs.NI ·2026-10-07 ·John Sengendo, Andreas Kassler, Fabrizio Granelli
published
Abstract · Accepted at IEEE GLOBECOM 2026

Future 6G networks are expected to operate as intelligent and autonomous ecosystems where monitoring, prediction, and control are integrated into continuous self-optimization loops. However, many digital-twin-based network management approaches still act mainly as synchronized replicas of the network. They observe the state, report degradation, and trigger corrective action only after performance risk has appeared. This leaves a gap between the vision of a cognitive digital twin (CDT) and the behavior of conventional reactive control loops. In this paper, we propose a Predictive Cognitive Digital Twin (PCDT) framework that closes the loop from observation to predictive cognition to proactive resource control. PCDT maintains a persistent traffic world model, forecasts near-future load, and allocates capacity for the predicted horizon peak before a Service Level Agreement (SLA) violation occurs. Evaluated on a real-world traffic trace, PCDT achieves the lowest allocation error among all benchmarked baseline frameworks, reducing MAE by 43% and RMSE by 32% relative to the best baseline (reactive control). Relative to the threshold heuristic, the only baseline with zero violations, PCDT reduces mean allocated capacity by 50%, reconfiguration churn by 56%, and total operating cost by 50%, indicating substantially more efficient performance. These results show the proposed framework advances the digital twin (DT) operation mechanism from a passive representation towards a cognitive, proactive control, and "intent-aware" mechanism aligned with autonomous 6G network vision.

0.45

ATLAS: Adaptive TDA-guided Landscape-Aware Transistor Sizing

cs.AR ·2026-10-07 ·Youngmin Oh, Jihwan Won, Yuntae Park, Bosun Hwang, Suwan Kim
code/data
Abstract

Analog transistor sizing, finding design parameters that simultaneously satisfy multiple performance specifications, is a labor-intensive bottleneck in circuit design. To support analog circuit experts, various automation methods have been proposed, including Bayesian optimization (BO), reinforcement learning (RL), and others. Yet existing methods are oblivious to the topological structure of a feasible design space, which can fragment into disconnected regions due to operating-regime transitions, conflicting specification trade-offs, and nonconvex device physics. This topological blindness causes the optimizer to converge within a single feasible region while missing others that may contain superior designs. To address this limitation, we propose ATLAS. a BO framework utilizing Topological Data Analysis (TDA). At each iteration, a Mapper graph is constructed over a surrogate-predicted feasible region to estimate connected regions, enabling topology-aware exploration from the very first iteration without any observed feasible points. A topological sensitivity score classifies candidates as bridge, frontier, or interior points, injecting a targeted exploration bonus into the acquisition function. Experiments on four analog circuit benchmarks in the GF180 and SKY130 processes demonstrate that \coin finds feasible designs with significantly fewer simulations than baselines, including RL and BO methods. To the best of our knowledge, this is the first work to apply topological data analysis to analog circuit design automation. The official implementation is publicly available on https://github.com/youngmin0oh/atlas.

0.44

LDPGraph: Locally Differentially Private Graph Synthesis by Exploiting Neighborhood Structure

cs.DB ·2026-10-07 ·Jiawei Dong, Zhikun Zhang, Quan Yuan, Zhe Liu, Yunjun Gao
top venuecode/data
Abstract · Accepted at IEEE ICDE 2027. Extended version with appendices

The widespread application of graph data inevitably brings significant privacy risks, as its unprotected use can lead to the leakage of sensitive information. These risks are particularly acute in the setting with an untrusted curator, where the data remains decentralized and each user only holds the connections to their neighbors. To mitigate such privacy risks, we adopt local differential privacy (LDP) to collect users' private information and generate a synthetic graph. However, existing methods suffer from either excessive noise injection by perturbing the local adjacency lists or significant structural information loss due to the simplistic graph encoding process. To address these issues, we propose LDPGraph, an effective graph synthesis algorithm that takes one step further by exploiting neighborhood structures under LDP. To obtain neighborhood statistics beyond degrees, LDPGraph aggregates a noisy global view from perturbed adjacency lists and combines it with projected local connections to estimate node-level triangle counts. To correct the structural inconsistency caused by separately perturbing degree and triangle count, LDPGraph jointly refines them into feasible degree-triangle targets. To reconstruct a global graph from these estimated targets, LDPGraph adopts a triangle-first strategy that first preserves local clustering structures and then fulfills remaining degree requirements. Extensive experiments on four real-world datasets and multiple commonly used graph metrics validate the superiority of LDPGraph. Source code is available at https://github.com/ZJU-TrustAID/LDPGraph.

0.44

Sufficient quantum provenance: retained fields and certified recording precision

cs.ET quant-ph ·2026-10-07 ·George Kourousias, Sergio Carrato, Roberto Pugliese, Aljoša Hafner, Francesco Guzzi +3 more
Abstract · v3: Revised abstract and introduction; clarified relation to prior work; added references. Scientific results unchanged

Quantum executions sharing a provenance record can produce different output distributions. We ask which execution details to retain, and at what precision, to guarantee agreement within a declared Hellinger tolerance. Under a specified channel model, we combine independent stochastic mechanisms and coherent rotation differences to certify recording precision without output simulation or explicit dependence on register dimension. For four-qubit quantum approximate optimization at depth one, a 13-bit comparison record suffices at tolerance 0.05 over a declared continuous noise range, uniformly in programmed angles for each fixed task. Pairwise separation constraints characterize sufficient field sets on finite context classes, exposing omissions that pass individually but fail jointly. Separately, six-qubit hardware measurements over six tested contexts certify minimum label counts of two for raw outputs and one after a fixed decoder, at tolerance 0.15 with 97\% simultaneous confidence under stationary independent shots. Execution records thereby carry checkable agreement guarantees.

0.41

CATune: Structural Constraint-Aware Bayesian Optimization for DBMS Configuration Tuning

cs.DB cs.LG ·2026-10-07 ·Fangping Lan, Qi Zhang, Eduard Dragut
Abstract · 14 pages including references, 10 figure, 2 tables. Accpeted by PVLDB, Volume 19, 2026

Modern DBMSs expose hundreds of configuration knobs, resulting in a high-dimensional and heterogeneous search space that makes automated tuning costly. Existing ML-based tuning systems typically treat the configuration domain as box-constrained and rely on workload feedback to implicitly capture inter-knob relationships. However, DBMS documentation specifies deterministic knob dependency constraints, particularly ordering constraints, that characterize structurally valid regions of the configuration space. We present CATune, a constraint-aware Bayesian optimization (BO) framework that models deterministic inter-knob ordering constraints as structural components of the search domain. Instead of learning feasibility boundaries through sampled violations, CATune performs optimization within a constraint-consistent subspace. We develop a topology-aware sampling strategy that respects dependency structure during exploration and avoids the inefficiencies of post-hoc constraint handling. To enable automated constraint discovery, we further design a precision-first extraction pipeline that combines LLM-based parsing with reliability safeguards to mitigate hallucinated dependencies. Experiments on PostgreSQL and MySQL using TPC-C and TPC-H workloads show that CATune substantially improves both sample efficiency and final tuning quality across surrogate models and BO frameworks. Under default ranges, CATune reaches the baseline optimum up to 12.5x faster and improves throughput by up to 63.37%. The improvements persist under knowledge-guided reduced ranges and alternative optimization implementations. These results demonstrate that explicitly modeling system-defined deterministic ordering constraints enhances optimization robustness and system stability.

0.41

ERP Event Logs: A Canonical Form for KPI Time-Series Extraction and Characterization

cs.DB ·2026-10-07 ·Sherri Hadian, Adrian Rebmann, Atacan Korkmaz, Gregor Berg, Ulf Brackmann
top venue
Abstract · accepted in D2AI at IEEE ICDM 2026. Version 2 corrects a mislabeled rate family in the forecastability figure (the "receipt rate" row is the…

Enterprise resource planning (ERP) systems record operations as transactions spread across hundreds of normalized relational tables. Extracting an operational key performance indicator (KPI) from this schema requires a separate, bespoke join for almost every metric. We show that an event log, a canonical (case, activity, timestamp) representation derived from the same tables, collapses this heterogeneity into one flat structure, from which KPI families reduce to reusable operations after a one-time mapping. Using standardized event logs built from hundreds of SAP S/4HANA customers under a shared activity ontology, we define three families of process-derived KPIs (volumes, durations, and rates) and extract them uniformly across customer systems for two common end-to-end organizational processes, order-to-cash and procure-to-pay. We normalize rates and screen series for minimum coverage to support comparison across customers of very different sizes. We then characterize heterogeneity across industries and tenants using descriptors derived directly from the canonical form, before analyzing the resulting KPI time series. The same process executes heterogeneously across customers and industries, yet this shared representation enables direct comparison of sales- and procurement-side signals, including lead times and order arrivals. Using this shared representation, we characterize which KPI families are forecastable and find a recurring ordering from volumes through durations to rates across both processes and a highly heterogeneous customer base; a pretrained time-series foundation model (Chronos) is competitive with classical baselines such as ARIMA and ETS on rate and duration series, while classical models retain an edge on volumes.

0.40

Analytical Modeling of Open- and Closed-Loop Dispersive Molecular Communication Channels with Pulsatile Flow

cs.ET physics.flu-dyn ·2026-10-07 ·Theofilos Symeonidis, Fardad Vakilipoor, Timo Jakumeit, Robert Schober, Nunzio Tuccitto +1 more
published
Abstract · 15 pages, 8 figures. This paper substantially extends the preliminary version presented at the 13th ACM International Conference on Nanoscal…

Molecular communication (MC) is a communication paradigm in which information is conveyed through the release, propagation, and reception of molecules. Many envisioned healthcare applications of MC are expected to operate inside the human body, where the cardiovascular system (CVS) may serve as the physical propagation environment and molecular transport is governed by diffusion and blood flow. Although blood flow is inherently pulsatile, most analytical MC channel models assume steady flow. In this paper, we develop a time-variant analytical model for dispersive MC channels with pulsatile flow. We derive the straight-duct response as a Normal distribution with time-variant mean and variance, capturing the combined effects of diffusion and pulsatile flow, and extend it to closed-loop channels through a wrapped-Normal representation. The model is validated against three-dimensional (3D) particle-based simulations (PBSs) for synthetic and physiologically motivated velocity waveforms. We further derive a closed-form first-order approximation for the first-arrival peak time and introduce the nondimensional indicator $S_{\mathrm{Rx}}$ for assessing the applicability of the reduced one-dimensional (1D) model. Our results show that pulsatility has the strongest influence when molecular transport is predominantly advective and the temporal flow variations are not averaged out, whereas stronger diffusion and faster pulsations reduce its impact on the received signal. Moreover, a PBS-based parameter sweep supports $S_{\mathrm{Rx}} \ge 2$ as a practical criterion for the applicability of the proposed analytical model. Finally, relating the analytical assumptions to representative blood-vessel classes shows that model applicability depends not only on the transport conditions but also on physiological properties such as vessel rigidity, geometric uniformity, and blood rheology.

0.39

Efficient Heuristics and Machine Learning Approach for Fault Characterization in Distributed Self-Stabilizing Programs

cs.DC ·2026-10-07 ·Amit Garu, Duong Nguyen
Abstract · 11 pages, 9 figures

Modern large-scale systems rely on distributed protocols to maximize efficiency while preserving the correctness guarantees of single-process execution. However, designing such protocols is non-trivial: adding resources to a system inherently increases its complexity, which in turn introduces faults that must be addressed during design. One such fault class, arising in systems that use distributed shared memory (e.g., replicated databases), is consistency violating fault (cvf), a fault in which a process accessing shared memory reads stale data previously written by another process. Cvfs are inherent to systems that prioritize availability over strict consistency. Prior work has shown that self-stabilizing programs remain correct under high availability, provided such faults stay within a tolerable bound. Characterizing the behavior of self-stabilizing programs in the presence of cvfs can therefore help system designers build systems that are both correct and efficient. In this paper, we study self-stabilizing programs under cvfs to understand when and where such faults occur, enabling runtime measures that improve computational efficiency. Since exhaustive state-space exploration does not scale, it suffers from state-space explosion; our study investigates two complementary approaches: a conflict-based state-space exploration technique for monotonically stabilizing $(Δ + 1)$-Graph Coloring programs, and a machine-learning-based approach for a Maximal Independent Set program. We propose a conflicting-edges-based state-space exploration algorithm alongside a rigorously validated ML model, evaluated on an out-of-distribution dataset through structural consistency checks against known graph invariants.

0.36

Perspective: Content-addressable memories as a computing primitive for today's AI and beyond

cs.ET ·2026-10-07 ·Paul-Philipp Manea, Bo Wen, Peiyi He, Can Li, John Paul Strachan
position
Abstract · 30 pages, 8 figures, 3 boxes. Perspective article, submitted to Nature Communications

Modern artificial intelligence is predominantly executed on computing architectures optimized for dense linear algebra. While this has enabled the success of contemporary neural networks and motivated compute-in-memory (CIM) architectures, a growing class of artificial intelligence (AI) workloads depends on associative retrieval, identifying stored information by content or similarity rather than by explicit memory addresses. Such operations are central to transformer attention, tree-based inference, genomic search, and other retrieval-intensive applications, yet remain inefficiently supported by conventional memory systems. In this Perspective, we argue that content-addressable memories (CAMs) provide a complementary hardware primitive for associative processing in AI. We review their ability to perform massively parallel in-memory matching, discuss how emerging memory technologies can improve density and energy efficiency, and identify hierarchical search, application-specific architectures, hardware-aware learning, and heterogeneous integration with CIM as key directions for scalable associative computing. Together, these developments suggest that associative retrieval should complement linear algebra as a foundational computing primitive for future AI systems.

0.34

A Distributed Accelerated Preconditioned ADMM for Solving Quadratically Regularized Optimal Transport on Bipartite Graphs

cs.DC math.OC ·2026-10-07 ·Jiayi Zhu, Haihang Lan
Abstract

Optimal transport (OT) has found broad applications in machine learning, economics, and management. Classical OT models, however, are not directly suited to allocation problems with graph-structured connectivity and unbalanced mass. Distributed OT formulations on bipartite graphs address these requirements but are typically solved by ADMM, whose slow convergence limits communication efficiency. We adapt the accelerated preconditioned ADMM (AP-ADMM) to the quadratically regularized distributed OT model and develop DAP-ADMM, a fully distributed algorithm in which each node communicates only with its graph neighbors. For fixed local penalty parameters, we establish global convergence and non-ergodic $O(1/k)$ bounds for both the KKT-residual norm and the primal objective gap. We derive an exact solver-free routine for the interval-constrained projections arising in the local ADMM subproblems. We propose a distributed self-adaptive penalty mechanism that updates node-specific penalties using locally computable KKT-residual components without global aggregation. Numerical experiments confirm the efficiency of the solver-free routine and the practical acceleration of DAP-ADMM over standard distributed ADMM.

0.32

Predictive Rerouting of Connected and Automated Vehicles Using Traffic and Charging Demand Forecasts

cs.NI eess.SY ·2026-10-07 ·Tony Kinchen, Ting Bai, Andreas A. Malikopoulos
hot direction
Abstract · 8 Pages

In this paper, we consider the problem of predictive rerouting of connected and automated vehicles (CAVs) in mixed traffic with electric vehicle charging demand. We provide a framework that combines a diffusion convolutional recurrent neural network (DCRNN) with a routing policy that accounts for congestion, CO2 emissions, route length, and charging demand. The DCRNN uses historical network observations to forecast traffic conditions and charging demand. These forecasts are then used to evaluate feasible alternative routes for eligible CAVs. A route change is accepted when the alternative preserves connectivity to the original destination and improves the prescribed route cost. We evaluate the proposed framework in SUMO under controlled traffic disruptions at five CAV penetration levels, ranging from 5% to 45%. We compare its performance with K-shortest-path routing, predictive-density routing, V2X proactive routing, and a reference scenario without rerouting. In the considered scenarios, the proposed framework reduces the average travel-time index by approximately 1.8% relative to the reference scenario and achieves the lowest average travel-time index, highest average speed, and lowest aggregate CO2 emissions among the active routing methods. Across all five penetration levels, it accepts 41 route changes, compared with 146 for K-shortest-path routing and 153 for V2X proactive routing. The reference scenario retains lower aggregate emissions and distance traveled, illustrating the tradeoff between congestion reduction and the additional travel associated with rerouting.

0.31

Communication-Aware Qubit Placement and Automatic Node-Count Allocation for Distributed State-Vector Simulation

cs.DC cs.ET ·2026-10-07 ·Íñigo Aréjula-Aísa, Sergio Iserte, Petter Sandås, Ricard S. Raigada-García, Antonio J. Peña
Abstract

Quantum circuit simulation enables the execution of large-scale quantum algorithms by partitioning qubits across multiple computational nodes. However, an inefficient qubit placement leads to excessive inter-node communication, limiting performance and scalability. This work introduces a framework that computes a communication-aware qubit placement offline, minimizing data exchange among nodes, and selects, before circuit execution, the largest node count whose performance degradation stays below a user-defined threshold. The framework is integrated with the QuEST simulator. Evaluations on synthetic benchmarks and real-world QASMBench circuits show that optimized placements reduce communication and achieve runtime speedups of up to 2.3x over the identity placement, particularly in circuits with sparse, structured interactions. Our results show that qubit placement strongly affects the efficiency of distributed simulation and provide a practical, resource-scalable solution for quantum workload

0.30

Reconciling Bottom-Up Metrics with Top-Down Reporting for Cloud Carbon Accounting

cs.DC ·2026-10-07 ·Philipp Wiesner, Loïc Lannelongue, Alexander Acker, Odej Kao
publishedsurveyposition
Abstract

Organizations seeking to reduce the carbon footprint of their cloud applications must rely on two largely disconnected ways of measuring it, each with important limitations. Bottom-up metrics like Software Carbon Intensity (SCI) provide signals for carbon-aware optimization, but tenants lack the information needed to account for many provider-side overheads. Top-down reports by cloud providers provide a more comprehensive view of a tenant's carbon footprint but are coarse and methodologically opaque. Because the two approaches differ in scope and reporting frequency, the current state of the art treats them as decoupled: bottom-up metrics for optimization, top-down reports for corporate disclosure. We argue that the two signals should be reconcilable. A per-workload metric whose improvements never surface in the provider's audited report will not be adopted for accountability at scale. We survey the state of the art in cloud carbon accounting and propose a vision for reconciled SCI (rSCI): a per-workload metric that re-anchors a bottom-up energy estimate to the cloud provider's top-down report through a residual. By decomposing and allocating this residual according to its physical drivers, rSCI can preserve operational incentives while also attributing idle capacity and embodied carbon to the workloads that drive them. We show that today's top-down reporting is still too coarse and methodologically inconsistent to support credible reconciliation and lay out concrete steps in provider reporting and standards to make it practical.

0.30

BenchmarkAnything: Agent-Driven Construction of Simulator-Ready Microarchitecture Benchmarks

cs.AR ·2026-10-07 ·Xiangfeng Sun, Ceyu Xu, Chen Bai, Yuan Xie
benchmarkcode/data
Abstract

The selection of benchmark workloads is of paramount importance in computer architecture, as it establishes the yardstick against which architectural innovations are measured and guided. Yet for decades, the SPEC benchmark suites, comprising merely tens of workloads, have been the de facto standard in academic architectural research, where they are frequently treated as a principal evaluation and optimization target. When a suite this small is relied upon so heavily, it risks architectural overfitting; as our research and prior studies demonstrate, an overly narrow focus can mislead design decisions by overvaluing certain innovations, producing cores that excel on SPEC benchmarks yet underperform on broader, realistic workloads. To mitigate this overfitting, adopting a large, comprehensive benchmark suite is the natural solution. However, the immense engineering effort required to strip software into the clean, interference-free binary executables demanded by simulators often makes this highly impractical. In this work, we demonstrate that AI agents provide an elegant solution to this challenge. Rather than manually curating yet another static benchmark suite, we introduce an agent-driven workflow capable of autonomously transforming arbitrary open-source repositories into simulator-ready executables. This automated approach makes workload collection highly scalable, allowing us to rapidly harvest hundreds of diverse applications from public repositories into our benchmark suite. Through a comparative analysis of our agent-generated suite against SPEC, we show that it not only achieves higher-fidelity performance assessments but also uncovers novel architectural insights that traditional, static suites fail to expose.

0.29

Performance Portable $\mathrm{SU}(N)$ Lattice Gauge Theory Simulation with Kokkos

cs.DC hep-lathep-phphysics.comp-ph ·2026-10-07 ·Wei Sun
Abstract · 14 pages, 5 tables, 4 figures

The increasing diversity of high performance computing systems makes separate, architecture specific implementations of lattice gauge theory algorithms costly to maintain. We present \texttt{kwqft}, a performance portable Kokkos implementation of Wilson pure gauge Monte Carlo simulation for $\mathrm{SU}(N)$ Yang-Mills theory in an arbitrary number of space-time dimensions. The gauge group order $N$ and the dimension $D$ are compile time parameters. A single source targets the Serial, OpenMP, CUDA, HIP, and SYCL execution spaces, with MPI halo exchange overlapped with interior updates. The implementation reproduces the exact two-dimensional plaquette and published three and four dimensional values for gauge groups up to $\mathrm{SU}(17)$. On an NVIDIA A100 the Kokkos CUDA backend is competitive with a native CUDA code, SIMD acceleration improves the OpenMP path on Armv9 processors, and a large scale speedup is demonstrated for an $\mathrm{SU}(4)$ lattice on the LineShine supercomputer, currently ranked first on the TOP500 list.

0.29

Implementation Guidelines for Data Quality Metrics

cs.DB cs.LG ·2026-10-07 ·Philipp Jung, Katinka Becker, Felix Biessmann, Valerie Restat, Martin Seyferth +2 more
code/data
Abstract · 14 pages, 7 figures, 5 tables. Submitted to EDBT 2027

Despite decades of data quality (DQ) research, a gap remains between DQ dimensions, such as accuracy or completeness, which the literature defines in textual form, and DQ tools, which typically implement low-level checks that are not aligned with these dimensions. ISO/IEC 25024 and ISO/IEC 5259 attempt to bridge this gap by defining DQ metrics for each dimension. However, these DQ metrics are hardly used, because the standards leave open how to implement them: for example, the metric for syntactic accuracy counts syntactically accurate values, but does not state how to decide that a value is syntactically accurate. This simply moves the problem to another level without solving it. As a result, DQ assessment currently cannot build on the standards. In this paper, we make the ISO DQ metrics executable. We classify all data-level metrics of both standards into (i) generalizable metrics that need no input beyond the data, (ii) parameterized metrics whose parameters can be learned from clean reference data or set by an expert, and (iii) non-generalizable metrics that need qualitative judgment and cannot be automated. For the metrics that can be automated, i.e., categories (i) and (ii), we propose implementation guidelines that resolve what the standards leave open. We realize the guidelines in dqmeasure, an open-source library of 20 metrics that learns these parameters from reference data instead of relying on manually defined rules. Our experiments on real-world and synthetic datasets show that the metric scores decrease monotonically with an increasing number of injected errors, decline together with downstream ML performance, and scale linearly with the number of rows, which enables automated DQ monitoring based on the standards.

0.28

Adaptive E2E Monitoring Path Selection with Deep Reinforcement Learning in Multi-domain Optical Networks

cs.NI ·2026-10-07 ·Soham Choudhury, Martin Bojinov, Jason P. Jue, Genya Ishigaki
published
Abstract · To be published in Journal of Optical Communications and Networking

Network slicing over multi-domain optical networks enables resource isolation for high-bandwidth, virtually-dedicated networks. However, guaranteeing end-to-end (E2E) performance remains challenging when domains withhold internal topology and performance metrics from external entities. Under this limited visibility, a control function (slice coordinator) needs to select E2E monitoring paths to detect and localize link failures. This paper formulates two complementary optimization problems, the initial and progressive monitoring path selection problems, which together realize closed-loop autonomous monitoring at the slice coordinator level. The initial problem maximizes failure coverage without prior information, while the progressive problem concentrates additional paths around suspected failure locations identified from prior monitoring. We propose a deep reinforcement learning (DRL) algorithm that adaptively selects an optimal set of monitoring paths for each phase. Experiments using the GNPy optical network simulator demonstrate that our approach outperforms baseline algorithms for both problems and achieves near-optimal localization in the initial selection, suggesting the feasibility of AI-driven closed-loop failure localization in future multi-domain optical network architectures.

0.27

Indexing Designs and Adaptive Data Distribution Optimization for In-Memory Databases on Hybrid DDR--CXL Memory

cs.DB ·2026-10-07 ·Huijie Cao, Shun Yang, Yinan Zhang, Huiqi Hu, Xuan Zhou +1 more
Abstract

CXL memory expands single-node memory capacity for in-memory databases but has higher latency and lower bandwidth than local DDR. Hybrid DDR--CXL databases require joint indexing and data distribution design. Indexing designs differ in access and migration paths and constrain the placement of indexes and tuples, creating trade-offs in runtime performance, DDR memory efficiency, and system integration complexity. Their performance impact is thus difficult to determine. Indexes and tuples may differ in access characteristics: placing them in the same tier may reduce DDR efficiency, while separate management adds tracking and decision overhead. We propose five indexing designs based on two approaches: treating the two tiers as a unified data space or managing them separately. We analyze their trade-offs and underlying causes and develop adaptive data distribution optimization for single-index designs in which the database explicitly manages distribution. The method manages index objects and tuples separately, reduces metadata overhead through hierarchical hotness tracking and an index object residency policy, and extends benefit-aware admission and randomized eviction to determine and adjust their placement according to each type's expected benefits and memory footprints. We implemented the designs in a hybrid DDR--CXL database prototype and evaluated them using YCSB. The single-index design with direct addressing achieves the best performance under most workloads and system configurations. The dual-index design performs better mainly when DDR capacity is limited or accesses are highly concentrated, and has relatively low integration complexity. Adaptive distribution optimization improves DDR memory efficiency and system performance, enabling single-index designs to achieve up to $1.59\times$ the throughput of their counterparts with non-separated management.

0.27

When Algorithmic Exploration Becomes Cheap: A Case Study of Agentic Research in EDA

cs.AR cs.AI ·2026-10-07 ·Keren Zhu, Yu Deng, Xiaoyu Hao, Liwen Jiang, Zijian Jiang +4 more
Abstract · 12 pages, 10 figures

As EDA researchers, we conducted eight deliberate trials of agentic algorithm exploration, selecting several topics outside our areas of depth. One faculty member and seven students participated, including students without publication experience. With limited intervention in the algorithms, agents developed mathematical constructions, analyzed existing tools, and implemented improvements; some efforts fell short of their practical goals. We also used AI to collect, classify, and analyze 8,420 papers from four EDA conferences and two journals over 2022-2026. Among 2,380 primary-core papers, we classified 97.7% from titles and abstracts as computationally closed, including work on new formulations. Together, these observations suggest that much of EDA offers an executable environment for increasingly accessible algorithm research. We see an opportunity for tool developers to investigate ideas they previously lacked time to pursue. We also ask how EDA should validate and reward research when results become easier to produce than to examine, and what papers and venue labels will continue to tell us about a contribution.

0.27

vLLM-Omni Technical Report: A Unified Serving Runtime for Omni-Modality Generation

cs.DC ·2026-10-07 ·vLLM-Omni Team
Abstract · 34 pages, 11 figures

Interaction with intelligent systems is expanding beyond text-centric chatbots and coding agents. Speech-native assistants, visual generation and editing, world-model environments, and robot action loops require models that emit text, audio, images, video, and actions. These models differ in execution pattern: multi-stage autoregressive omni and TTS pipelines, iterative diffusion or flow-matching generators, and longer-lived world-model or robot loops that carry state across steps. As a result, serving is no longer a single text decode loop, but a heterogeneous multi-stage workflow with cross-stage transfer, streaming, and session-shaped interaction. Existing inference stacks are typically optimized for one architecture family. LLM servers deepen autoregressive scheduling and KV management, while diffusion stacks deepen denoising and parallel generation. Neither provides a shared control plane for pipelines that emit speech, pixels, or actions through separate generators, so production deployments often fall back to ad-hoc composition across disjoint runtimes. We present vLLM-Omni, a unified serving runtime for omni-modality generation. vLLM-Omni organizes each workload as a multi-stage pipeline under a single orchestrator that admits requests, advances them across stages, and demultiplexes streaming outputs. Specialized engines and stage replicas provide compute; a connector carries heavy payloads on the data plane; and session-oriented control supports long-lived duplex, world-model, and robot workloads. This report covers the architecture (stage-level KV paths, replica pools, multi-hardware platforms, and efficiency stack) and OpenAI-compatible and OpenPI APIs for omni, TTS, image/video, world-model, robot, and duplex workloads. We evaluate on the multimodal nightly CI on H100 (TTS and MiniCPM-o on H200), focused on Qwen3-Omni.

0.27

A Compositional Perspective on Communication-Control Co-Design for Mobile Broadband Systems Beyond 6G

cs.NI ·2026-10-07 ·Angelo Vera-Rivera, Atefeh Termehchi, Ekram Hossain
Abstract

Anticipated applications of beyond sixth-generation (B6G) mobile broadband networks will require the co-design of communication and control subsystems within the network architecture. Existing co-design approaches are predominantly optimization-driven, integrating subsystems through joint optimization problems. While this approach is effective in relatively simple systems, such formulations present challenges in (i) modular subsystem representations, (ii) tracing the propagation of requirements across subsystems, and (iii) systematically analyzing design-space feasibility, particularly as the number and complexity of interacting subsystems increase. In this article, we present a compositional perspective on co-design based on the formal theory of co-design. We examine how the notion of composition introduced by this theory can address the challenges of the conventional optimization-driven approach. Building on this perspective, we propose a composition-driven methodology for communication-control co-design in B6G networks and illustrate it through a wireless-assisted robotic control case-study. We then discuss how the composition-driven and optimization-driven co-design approaches can complement each other and why this complementarity may be beneficial for B6G networks. Finally, we identify key research challenges and future directions toward the practical adoption of the proposed methodology.

0.25

Reproducible LLM Inference Benchmarking: A Sequential Isolation Protocol for Regression Testing

cs.PF cs.LG ·2026-10-07 ·Arnold Olympio, Juan Manuel Servera Bondroit, Wael Abdelmalek, Guang Lu, João Carvalho
benchmarkcode/data
Abstract

Reproducible benchmarking of Large Language Model (LLM) inference is challenging because repeated measurements can vary with execution and system state. We present the Sequential Isolation Methodology, a controlled benchmarking and regression-testing protocol designed to reduce between-run measurement variance while deliberately varying workload concurrency. We evaluate three representative open-source LLMs on an NVIDIA A100 80GB GPU using vLLM 0.9.1 across six context sizes and eight concurrency levels, with five repetitions per configuration. The final protocol reduces average coefficient of variation (CV) from 15.2% in the least controlled methodology stage to 2.2% under the final protocol; using CV computed across the five repetition-level median (P50) TTFT values per configuration, 113 of 144 configurations (78.5%) achieve CV below 3%. The measurements also show a marked latency transition between 200 and 500 concurrent users on the tested stack and descriptive differences in P99 latency across the three models. We additionally provide an explicit cost break-even model with sensitivity to API pricing. The protocol is intended to provide a stable reference for reproducible comparison and regression testing rather than to predict absolute behavior under uncontrolled production traffic. Infrastructure-as-Code and benchmark scripts support replication of the experimental environment.

0.25

Fast and Memory Efficient Offload Training Framework with Hybrid XPU Computation

cs.DC ·2026-10-07 ·Zhiyi Yao, Zuning Liang, Yuedong Xu, Jin Zhao, Jessie Hui Wang +1 more
Abstract

With the ever-growing size of deep learning models, GPU memory is prone to being insufficient during training. A prominent approach is ZeRO-Offload, which moves the optimizer states to CPU memory and performs parameter update using CPU. However, the deficiencies of ZeRO-Offload include low GPU utilization, imperfect overlapping of communication and computation, and inflexible offloading. In this paper, we leverage Direct Host Access (DHA) on the GPU that can compute data in CPU memory, forming a novel hybrid on-GPU and DHA. We design and implement MemFerry consisting of an execution scheduler and a shadow model. The scheduler strategically chooses layers of parameters for DHA computation and transmits the remaining parameters to GPU memory simultaneously to shorten forward propagation time, and further loads DHA parameters to GPU memory to reduce backward propagation time. The shadow model presents a unified memory abstraction for the parameter partitions stored separately in GPU and CPU memories. To further reduce GPU memory usage, we present MemFerry along with its dynamic programming algorithm that offloads gradients to CPU memory via DHA. We further extend MemFerry to emerging scale-up domains with ScaleUp-MemFerry, which exploits otherwise underutilized accelerator interconnect bandwidth to assist host-to-accelerator data movement through adaptive multi-path transfer. Our experiments show that \system trains up to $1.68\times$ faster and MemFerry can train $1.52\times$ larger model compared to ZeRO-Offload on a single GPU, and increase training speed by at least $28.1\%$ when scaling to data parallelism on 8 GPUs. We further extend the design to a Huawei CloudMatrix384 scale-Up node with up to 8 NPUs, and our ScaleUp-MemFerry reduces the end-to-end iteration time by up to $20.7\%$ over DeepSpeed.

0.25

Age of Information in Queueing Systems with Merging Server Streams

cs.NI ·2026-10-07 ·Lucrezia Rossi, Leonardo Badia, Andrea Munari
Abstract

We present a novel analytical framework for evaluating the age of information (AoI) in multi-path queueing topologies where output streams from multiple servers converge into a single pipeline. This theoretical scenario finds immediate application in ultra-reliable low-latency networks and edge computing architectures, where multi-path routing and stream merging can be vital strategies for maintaining fresh status updates. Specifically, we investigate two setups: a split-merge network and a duplicate-merge network. The split-merge case has been addressed in prior work, but we show that existing studies are incorrect, and our approach provides instead an exact analytical formulation. Moreover, building on similar reasoning, we provide the first formal analysis of the previously unexplored duplicate-merge scenario. Finally, we show how to leverage our theoretical results to solve key system-level optimization problems, deriving the optimal randomized routing probabilities and traffic injection rates that minimize average AoI.

0.24

No Trace, No Claim: Two Contracts for Database Agents

cs.DB ·2026-10-07 ·Xiaofei Zhang
position
Abstract

LLM agents can generate database operations and explain their results, but current interfaces often leave a gap between generated plans, execution conditions, and claims presented to users. We argue that agent-facing data systems need two enforceable contracts. A plan contract defines what an agent may execute and reference; an evidence contract records the belief state, completeness, and provenance under which a result supports a claim. We instantiate these contracts in TGMS, a bi-temporal graph system in which an LLM plans over a fixed temporal operator interface. A static verifier checks plans before execution, and a claim verifier checks typed claims against content-addressed traces. Live model runs exposed two failures missed by input-only schemas and value-only grounding: nonexistent result fields and page-local counts reported as complete-result counts. Result-field checking makes the first a repairable rejection. Completeness propagation detects the second in all 15 controlled cases and misses all 15 when disabled. On the frozen CollegeMsg workload, TGMS reaches 0.408 typed-answer accuracy versus 0.064--0.284 for the evaluated baselines. On Bitcoin-OTC, direct SQL over the same bi-temporal store matches TGMS, showing no universal accuracy advantage for the fixed operator interface. On correction probes, TGMS and bi-temporal SQL answer historical-belief questions, while latest-state baselines cannot. Before claim gating, 21 of 220 answers contain an unsupported gated claim; after gating, none of 199 emitted answers does, at the cost of 21 fewer answers. The plan contract makes invalid plans rejectable and execution reproducible under recorded conditions, while the evidence contract makes gated claims faithful to cited evidence. Neither guarantees correct interpretation of user intent.

0.24

RFChipAgent: Multi-Agentic AI Flow for Analog/RF Chip Design

cs.AR cs.AIcs.LGcs.MA+1 ·2026-10-07 ·Awani Khodkumbhe, Yunfei Feng, Raj Rangarajan, Kevin Wang, Kamal Sahota
Abstract · 6 pages, 4 figures. Submitted to ACM/IEEE for possible publication

Analog/RF circuits remain the critical interface between digital computation and the physical world, and emerging standards from Wi-Fi 7 to 6G place stringent demands on them, yet analog/RF design remains one of the most labor-intensive steps in chip development. We present RFChipAgent, a first-of-its-kind multi-agent flow of large language model (LLM) agents for end-to-end analog/RF circuit design automation, in which AI agents collaboratively orchestrate the complete design flow under human supervision. RFChipAgent is built around four technical pillars. First, a multimodal retrieval-augmented generation (RAG) subsystem with private per-document FAISS indexing extracts design knowledge from existing engineering documentation. Second, a topology agent drives topology selection, and a schematic and testbench agent automates circuit and testbench assembly. Third, a closed-loop hybrid circuit-sizing engine combines Tree-structured Parzen Estimator (TPE) and CMA-ES optimization, evaluating every candidate in a simulator-in-the-loop framework. Fourth, a trust-scored simulation database accumulates verified performance data and builds an adaptive optimization model that informs subsequent trials. We validate RFChipAgent on a family of GF22FDSOI 60 GHz wideband mm-wave low-noise amplifier (LNA) topologies, demonstrating automated topology generation, specification-driven design-space exploration, and simulator-guided optimization. Experimental results show substantial reductions in design effort while maintaining signoff-quality verification. This work establishes a foundation for LLM-driven multi-agent electronic design automation (EDA) for analog/RF circuits.

0.23

LOCAA: An Agentic System for Automated Lossy Compressor Tuning

cs.DC ·2026-10-07 ·Khondoker Mirazul Mumenin, Dong Dai, Sheng Di, Franck Cappello, Jinzhen Wang
Abstract

Large-scale scientific simulations generate substantial data volumes, making lossy compression essential for reducing storage and data movement costs. However, users configure compressors through numerical error bounds (EBs) while often evaluating results using quality metrics and end-to-end performance. Because the relationship between an EB and these outcomes varies across datasets and compressors, identifying a suitable configuration typically requires exhaustive search, which can be time-consuming and computationally demanding. We present LOCAA, an Large language model-based scientific lossy compression auto-tuning agent that performs compression-in-the-loop search using tool-integrated execution, compressor-aware guidance, and persistent memory. LOCAA supports user-defined objectives and constraints without requiring a specialized search strategy. We evaluate LOCAA across three scientific applications, five compressors, and three use cases: fixed-ratio compression, compression tuning under multiple quality constraints, and compression tuning under quality and time constraints. For fixed-ratio search across twelve fields from two applications, LOCAA requires 1.98x fewer evaluation trials than binary search and 5.03x fewer than FRaZ on average. For compression ratio maximization under joint Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure constraints across six fields from the Community Earth System Model application, LOCAA reduces the average evaluation trials from 61.5 using binary search to 17, corresponding to a 72.4% reduction. Persistent memory further reduces the average number of trials by 27.9% across multiple timesteps of the same field. These results demonstrate the potential of tool-augmented LLM agents to provide flexible and efficient compressor tuning across diverse scientific data and user-defined objectives.

0.22

SUSpMV: A High Frequency Sparse Matrix Vector Multiplier on HBM Enabled FPGA written in SUS

cs.AR cs.DC ·2026-10-07 ·Lennart Van Hirtum, David Volz, Andreas Koch, Christian Plessl
published
Abstract · Accepted at the International Conference on Field-Programmable Technology (FPT) 2026. 9 pages, 9 figures

SUSpMV is a Sparse Matrix Vector multiplication (SpMV) accelerator written in the upcoming HDL SUS. By leveraging SUS's unique latency counting and inference mechanism, SUSpMV could be designed with very deep pipelines yet small design complexity overhead. This enables an efficient implementation which employs all 32 HBM channels on the Alveo U280 FPGA at 400MHz for streaming matrix data into 32 Compute Units (CUs). Input and output vectors are stored in DDR memory, which allows zero-overhead chaining of multiplications and increases overall system memory bandwidth by not sharing HBM bandwidth with the CUs. A CU processes the SpMV in tiles of width 1024 and a dynamically chosen height, up to 32768. Each CU is capable of accumulating the multiplications with up to 6 separate matrix entries per cycle, resulting in the combined theoretical peak computational throughput of 153.6 GFLOPs. The matrix storage format is designed to exploit density variation within a given matrix by dynamically switching between one representation optimized for denser regions, and a second optimized for sparser regions, allowing jumps of up to 255 rows between each entry. Evaluation demonstrates a 79% geometric mean improvement over prior work on the same platform. We achieve a peak throughput of 144.9 GFLOPs, or 94% of our theoretical computational throughput, compared to 98 GFLOPs reached by prior work.

0.22

Behind a Simple Read: Understanding Work and Waiting in the Linux I/O Stack

cs.OS ·2026-10-07 ·Yang Shen, Wenzhe Zhang, Yong Dong, Min Xie, Wei Zhang +2 more
Abstract · 30 pages, 15 figures, 3 tables

A simple read interface provides uniform semantics but unpredictable performance. Using synchronous large-buffer Linux reads, we decompose the buffered-read path into work volume, processing time, and critical-path exposure. Reducing local cost or accelerating request submission need not shorten a complete read. Linux uses large folios to reduce metadata work and I/O-copy overlap to hide most cold-read copying; an experimental kernel prototype shows that more aggressive overlap and parallelism can shorten the host path when these mechanisms fail. Device waiting dominates SSD cold reads. We abstract first-completion wait F and subsequent completion capacity B_CQ from request-batch timelines. Real-SSD experiments characterize their variation with request and batch size, while MQSim traces explain their internal origins. A wait chain along the folio-bio-request mapping connects device completion to upper-layer waiting. Under strict batching, the median absolute cross-layer prediction error is at most 7.43%. Finally, we apply the completion model and wait chain to request supply, Linux readahead, data layout, and waiting policies. Cases cover gains, no gains, and reversals, comparing predictions with interventions to close the loop from observation and modeling to understanding and control and provide a measurable basis for layered I/O decisions.

0.22

Deadline-Aware Multi-Agent Reinforcement Learning for TSN-Based Vehicular Edge Networks

cs.NI cs.AI ·2026-10-07 ·Bernardo A. C. Pereira, Marcos Carvalho, Fatih Temiz, Shavbo Salehi, Melike Erol-Kantarci +3 more
Abstract

Vehicular edge computing (VEC) enables latency-sensitive applications by bringing computing and networking resources closer to vehicles. However, existing approaches often overlook network contention among co-located services with heterogeneous and dynamic latency requirements. While time-sensitive networking (TSN) provides bounded-latency communication, conventional and reinforcement learning-based schedulers struggle to adapt to highly dynamic vehicular environments and inter-queue dependencies. To address these limitations, we propose a multi-agent reinforcement learning (MARL) approach for queue-level scheduling in TSN-enabled VEC. Each TSN queue is assigned an autonomous agent that jointly learns the queue service order and time-slot duration to minimize deadline misses under speed-dependent latency requirements. We employ multi-agent proximal policy optimization (MAPPO) to enable coordinated yet autonomous scheduling decisions. Evaluation against single-agent, multi-agent, and non-learning-based baselines shows that MAPPO provides robust performance across different traffic profiles. Compared with centralized single-agent methods, it reduces service latency by up to 66.2% and improves reliability by up to 271.8%. Furthermore, unlike urgency-based heuristics, MAPPO ensures balanced scheduling while achieving lower inference times compared to other MARL methods.

0.21

Rendezvous under Variable Disorientation:The Algorithmic Power of Fixed Unit Distance

cs.DC ·2026-10-07 ·Koichi Wada, Toshimitsu Masuzawa, Yuichi Sudo
theory
Abstract

We study deterministic Rendezvous of two robots with lights under weak geometric and synchronization assumptions. Our focus is on Variable Disorientation (VD), where each robot may arbitrarily rotate, reflect, and rescale its local coordinate system at every \Look, and on VD+Fixed-Unit-Distance (FUD), where each robot's local unit distance remains fixed. We consider the one-sided visibility models FSTA and FCOM, and schedulers ranging from the energy-restricted RSYNCH and R-RSYNCH to fully asynchronous executions. Under VD, we prove color-independent impossibility results for FSTA under SSYNCH and for FCOM under ASYNCH, and establish sharp two-color boundaries under RSYNCH and R-RSYNCH. Under VD+FUD, we introduce periodic distance classes, which turn multipl icative changes of the physical distance into predictable cyclic phase shifts. This technique yields a two-color self-stabilizing non-$L$-Rendezvous algorithm for both FSTA and FCOM under RSYNCH and R-RSYNCH even with Non-Rigid movement. Under Rigid movement, periodic distance classes further yield an optimal three-color non-quasi-self-stabilizing algorithm for FSTA under SSYNCH, a four-color FSTA algorithm under $M$-atomic ASYNCH, and a four-color FCOM algorithm under CM-atomic ASYNCH. By guarding the latter phase mechanism with two additional colors and replacing its initial swap by passive initialization, we obtain a six-color FCOM algorithm under full ASYNCH, improving the previous twelve-color upper bound by a factor of two. These results show that a fixed local metric can serve as persistent cyclic memory, and clarify how metric stability, one-sided visibility, and scheduler atomicity jointly determine Rendezvous solvability and color complexity.

0.20

HPC-MQBench: Qualification-First Benchmarking on Slurm with a Single-Broker Kafka Evaluation

cs.DC ·2026-10-07 ·Sepehr Mahmoodian, Julian Kunkel
benchmark
Abstract · 8 pages, 4 figures, 3 tables

Messaging experiments on high-performance computing clusters must coordinate services, clients, and measurement within scheduler allocations. We present HPC-MQBench, a Slurm-orchestrated benchmark that links resumable experiment control to record checks, delivery accounting, and qualification before rate ranking. Its Kafka evaluation placed producers/controller, one broker, consumers, and monitoring on four nodes. With memory-backed logs, 120 workload configurations yielded 99 qualified observations, 14 outside the producer-delivery policy, and seven with invalid evidence. Two validation stages each repeated ten workloads in five blocks. In the final stage, the selected workload qualified in all five observations, with medians of 2,927 mebibytes per second balanced endpoint rate, 1.24 percent pending deliveries, and 2.56 seconds for the 99th-percentile latency. It qualified in only two observations in the earlier stage, where qualification improved after an allocation boundary without establishing a causal allocation effect. Selection is conditional on the stage and delivery policy. Resource measurements did not isolate a unique bottleneck. The contribution is a framework for distributed experiment control and auditable configuration selection at a fixed broker count; multi-broker support and scaling experiments remain future work.

0.18

Democratizing MoE inference on commodity GPUs with CoMoE

cs.DC cs.LG ·2026-10-07 ·Ruwen Fan, Yuezhi Zu, Minhui Xie, Junru Li, Qingda Hu +4 more
Abstract

Deploying Mixture-of-Experts (MoE) models relies heavily on Expert Parallelism, which generates intense inter-GPU communication. Consequently, state-of-the-art inference systems require high-bandwidth, P2P interconnects (e.g., NVLink) in datacenter GPUs to handle massive token routing, making deployment prohibitively expensive. Consumer GPUs offer comparable compute power at significantly lower cost, promising to democratize MoE inference for individuals and enable privacy-preserving local deployments. However, their bandwidth-limited (only weak PCIe bus bandwidth) and host-mediated interconnects (no P2P support) introduce severe communication bottlenecks. We present CoMoE, a communication-efficient MoE inference system that resolves this mismatch through novel host-centric routing. Our key insight is that the unique communication topology provides the opportunity to elevate the host to an active routing hub, which can fundamentally reduce communication volume and eliminate global synchronization-induced stalls. Specifically, for token dispatch, we introduce host-backed token multicast to write shared tokens to the host exactly once, eliminating outbound transmission redundancy. For token combine, we propose a fine-grained, token-level aggregation mechanism using host staging buffers, which replaces rigid global synchronization and mitigates straggler effects. Evaluation on RTX 5090 GPUs shows that CoMoE improves inference throughput by up to 1.46x, approaching the performance of NVLink-capable A800 GPUs at only 23.4% of the hardware cost.

0.17

Adaptive Semantic Communication with Residual Quantization for Resilient Vehicular Networks

cs.NI ·2026-10-07 ·Xuanhao Luo, Ruichen Gao, Sandeep Sreenivasan, Zhizhen Li, Yuchen Liu
Abstract

Vehicular networks require timely visual information exchange to support safety-critical applications such as cooperative perception and hazard awareness, yet transmitting high-dimensional camera data over bandwidth-limited and dynamic wireless links remains challenging. In this paper, we propose an adaptive semantic communication architecture for vehicular multicast that adjusts the amount of visual information transmitted according to current channel conditions, available communication resources, and latency requirements. Specifically, the visual encoder uses residual quantization to represent each image at multiple bitrate and reconstruction-quality levels. At runtime, the system selects an appropriate level based on the communication conditions, enabling it to trade visual fidelity for lower transmission overhead when network resources are limited. Context-adaptive entropy coding further compresses the quantized representations using offline-learned statistics, without requiring retraining for different channel conditions. This design allows the visual representation to be learned offline while the communication strategy is adapted online. Experimental results demonstrate competitive rate-distortion performance at ultra-low bitrates, improved delivery reliability and deadline satisfaction across diverse vehicular scenarios, and strong downstream perception performance, achieving up to 0.88 mAP50 while dynamically adapting to changing communication conditions.

0.15

An Interleaved Parallel Dependent Quantization Hardware Architecture for H.266/VVC

cs.AR ·2026-10-07 ·Jun Zhang, Weizhi Bian, Hao Zhang
Abstract

While dependent quantization in H.266/VVC delivers a high compression ratio, its strong serial nature and high complexity result in poor real-time performance, making it difficult to deploy in practical scenarios. To improve the real-time performance of dependent quantization with minimal degradation to its compression performance, we propose a interleaved parallel dependent quantization hardware architecture with low BDBR loss, which achieves four-channel parallel dependent quantization by time-division multiplexing most combinational logic. This architecture adopts the proposed intra-CG context simplification scheme and the encoding scheme that skips decAbslevel during rate estimation. The proposed design incurs a BDBR loss of only 0.42% under the All Intra configuration and 0.38% under the Random Access configuration, respectively. Implemented in Verilog HDL, the dependent quantization hardware architecture achieves quantization speeds of 4K@33.2, 91.4, 331.2, and 454.1 fps at QP = 22, 27, 32, and 37, respectively, when implemented on the Xilinx XCZU19 FPGA. When implemented on ASIC using the TSMC 28nm process standard cell library, the corresponding quantization speeds reach 4K@84.0, 231.0, 837.0, and 1147.7 fps.

0.13

QCATS: Query Context-Aware Transformer Slicing for Efficient Predictive Query Processing

cs.DB cs.AI ·2026-10-07 ·Yueying Li, Zhongle Xie, Ke Chen, Lidan Shou
Abstract

In-database predictive query processing increasingly applies Transformer-based models within relational pipelines. However, existing in-database inference typically exposes only tuple-level model inputs to the inference runtime, leaving relational predicates and metadata statistics invisible to neural execution planning. In this paper, we propose QCATS, a query context-aware transformer slicing framework that enables efficient sparse inference inside database systems. QCATS executes at query granularity: instead of routing individual tokens or tuples during inference, it uses query predicates and metadata statistics to pre-select context-aligned FFN slices before model execution. The framework comprises offline expert construction and lightweight query-level routing that dynamically selects experts during execution. QCATS further introduces system optimizations, including asynchronous CPU-GPU pipelines and routing-aware batching. Experiments on four predictive-query workloads with BERT-base and Qwen-0.6B show that QCATS achieves up to 4.42x latency reduction while preserving prediction accuracy comparable to dense baselines.

0.13

The Missing Fourth Term for the Emulation Tensor Memory Equilibrium (TME) Model: The Residue Deconstruction Cost

cs.AR cs.AIcs.PF ·2026-10-07 ·Harun Bayraktar, John Gunnels, Peter Caday
Abstract

The Tensor-Memory Equilibrium (TME) model of "FP8 is All You Need (Part 1)" calculates the execution time of Ozaki Scheme II emulation of fp64 as the maximum of a tensor-core term and a High-Bandwidth Memory (HBM) traffic term, plus a per-output reconstruction term. However, it omits the per-input deconstruction cost: every streamed fp64 operand must be scaled, rounded, and reduced modulo each of the $r$ moduli on SIMT pipes before any matrix multiply can issue. In this note we add this fourth term, calibrate its constant from the cuBLAS emulation path, and derive a closed-form operational-intensity threshold $\mathrm{OI}^{*} = c_q r P_{\mathrm{fp64}}/(8P_{\mathrm{int}})$ below which emulation cannot match native fp64 regardless of tensor-core throughput. On the NVIDIA B300 GPU the threshold is $\mathrm{OI}^{*}\approx 0.56$ FLOP/B. As a result, GEMV, SpMV, and low-batch GEMV, which are the memory-bound kernels the original paper claims to accelerate, are limited to 0.3-0.9x of native performance, and the 7-point stencil to 1.8x rather than the claimed 3.1x. Dense GEMM is unaffected as expected. We also show that precomputing and storing the residues moves the same cost into the bandwidth term, and we state the instruction count that an implementation would have to achieve to invalidate the bound.

0.08

Scaling Quantum Optimization to the Thousand-Qubit Scale with Distributed Quantum Sampling

cs.ET quant-phmath.OC ·2026-10-07 ·Amana Liaqat, Ahmed Darwish, Stephen DiAdamo, Dan Holme, Kieran McDowall +4 more
benchmark
Abstract

The Minimum Birkhoff Decomposition (MBD) seeks a sparse weighted sum of matchings and is a challenging optimization problem with applications in network scheduling and energy trading. We scale a quantum-assisted decomposition method by combining single-layer QAOA sampling with Extended Fully-Corrective Frank-Wolfe (E-FCFW) optimization, spectral graph partitioning, and greedy feasibility repair. We demonstrate the pipeline on the 1,354-node PEGASE bus test case (representing a 1,354-bus transmission grid with 1,710 transmission lines), whose 1,710 edges define a native 1,710-variable matching optimization problem requiring 1,710 qubits in the unpartitioned encoding. Partitioning enables distributed execution on IBM superconducting quantum processors. The best reported hardware result uses a 50-qubit partition bound, while reducing partitions to 20 qubits degrades convergence in the partition-size comparison. Larger partitions are more demanding for matrix product state (MPS) simulation, and bond-dimension tests show that truncating quantum correlations reduces candidate quality. The results therefore motivate retaining a substantive quantum sampling task within each partition as the overall problem scales. On the PEGASE-1354 benchmark, repaired QAOA samples achieve lower residual decomposition error than both simulated annealing and uniform random sampling baselines, while spatial circuit packing reduces hardware execution time by approximately a factor of three. These results demonstrate that combining spectral partitioning, spatial circuit packing, and classical feasibility repair provides an executable path for deploying gate-based quantum sampling on thousand-variable constrained optimization problems on current hardware.

0.07

Software & Programming 32

Agent4RE: A Self-Refining Multi-agent Framework for End-to-End Software Requirements Engineering and Benchmarking

cs.SE cs.AI ·2026-10-07 ·Yongjian Tang, Linhan Li, Thomas Runkler
benchmarkdatasetcode/data
Abstract · Accepted to the ASE@POVC track; The E2E requirements engineering benchmark is available https://zenodo.org/records/22815961

Existing LLM-based approaches for software Requirements Engineering (RE) typically rely on basic prompting strategies or rudimentary agent collaboration, under-utilizing the full potential of multi-agent systems. Meanwhile, available datasets focus on isolated subtasks, such as requirements extraction, classification, and completeness detection, leaving the absence of an end-to-end RE benchmark that spans from requirements elicitation to generation. We present Agent4RE - a self-refining multi-agent RE system that orchestrates specialized agents and incorporates two iterative improvement loops. To support evaluation, we construct RE-E2E - a real-world dataset built from human-written requirement specifications, enabling end-to-end assessment of RE workflows. Building on this foundation, we further propose two enhanced Agent4RE versions that incorporate either autonomous self-refinement or structured human feedback, and analyze their strengths and limitations across different scenarios. Evaluation on 8 Large Language Models (LLMs) demonstrates that all three Agent4RE variants consistently outperform a domain-context-augmented prompting baseline by average 8% in text-based metrics. The two enhanced variants achieve the highest LLM-as-a-judge and human ratings, surpassing two RE baselines by approximately 0.8 points on a four-point scale. This consistent performance establishes Agent4RE as a practical end-to-end RE solution for industrial environments.

0.57

Two-Variable Logic with Arbitrarily Many Successor Relations

cs.LO ·2026-10-07 ·Jakub Michaliszyn, Piotr Witkowski
theory
Abstract

We study two-variable first-order logic with a finite, input-dependent number of distinguished successor relations and arbitrary additional unary and binary predicates. Each distinguished relation must be the exact immediate-successor relation of some linear order on the common domain; the orders themselves are not available in the language and may have arbitrary order types. We prove that satisfiability belongs to 2NEXPTIME, uniformly in the number of successors. The proof gives finite certificates for possibly infinite models. A closure operation separates local neighbourhood descriptions, called star types, into those of bounded multiplicity and those that can be realized infinitely often. The bounded part is represented explicitly. Outside it, integer height vectors allow overlapping successor requirements to be matched without creating cycles or unintended identifications. Ordering the resulting path components densely then recovers exact successor relations. Consistent assignments of the additional binary predicates handle witnesses involving the bounded part. The same certificates decide infinite satisfiability in 2NEXPTIME and give a doubly exponential threshold above which a finite model guarantees an infinite model. A doubly exponentially large finite core can be forced with only two successors and unary predicates. For finite models, we prove that satisfiability subject to a bound on the total number of reference adjacencies lost by the other orders is NEXPTIME-complete, even when the budget is encoded in binary and the number of successors is part of the input.

0.50

Who Broke Me? Execution-Guided Repair of Behavioral Dependency Breaks

cs.SE ·2026-10-07 ·Juntao Chen, Tingting Bi, Yanlin Wang, Patanamon Thongtanunam
benchmark
Abstract · 21 pages, 4 figures, 12 tables

Dependency upgrades can break downstream projects without changing the library interface. Such behavioral breaking changes are difficult for developers to fix, because the failing test does not always point to the root API, the upstream API that causes the break. Existing LLM-based repair methods obtain evidence for the repair from compiler feedback or library documentation. However, a behavioral break produces no compiler feedback and is often undocumented. Agents that receive no upgrade evidence also usually do not retrieve upstream evidence themselves, and most of their failed repairs do not identify the root API. The library diff provides useful upstream evidence, but the root API must first be identified to select the relevant part of the diff. We present BBCFixer, a repair method that runs the failing test under the old and new library versions, ranks the calls whose return value differs to identify a candidate root API, and filters the library diff with that candidate. To evaluate repairs of behavioral breaks, we further introduce BBCBench, a benchmark of 100 behavioral dependency breaks in Python and JavaScript. We compare BBCFixer on BBCBench with two baselines: (1) Pure Agent, an agent that receives no upgrade evidence, and (2) BDUpdater, a method that mines library documentation. BBCFixer raises the average pass rate by up to 17% relative to Pure Agent and needs up to 27% fewer agent steps per successful repair. BBCFixer thus provides a way to select the upstream evidence behind a behavioral break, so that an LLM agent can repair behavioral breaking changes more effectively and efficiently.

0.47

A Chat Assistant for Software Exploration in a 3D Software Visualization

cs.SE ·2026-10-07 ·Malte Hansen, Karim Issa, Wilhelm Hasselbring
top venue
Abstract · 11 pages, 5 figures, accepted at VISSOFT 2026

We present a chat assistant for interactive software exploration, embedded in the 3D software visualization tool ExplorViz. The assistant builds upon current Large Language Models (LLMs) and enables users to ask questions about the currently visualized software system and trigger actions that change the visualization through natural language. We integrate the chat assistant in our ExplorViz frontend using the CopilotKit libraries such that probabilistic LLMs are combined with deterministic and tool-based actions similar to implementations using the Model Context Protocol (MCP). The chat assistant is also enabled to restructure the software system by adding, removing, or modifying parts of the software system in the visualization. An empirical experiment with eleven participants evaluated both perceived comprehension support and the tool calls that were triggered by the chat assistant. Participants rated the assistant's generated summaries and explanations as largely correct. They also reported high usability for actions like highlighting entities in the visualization and the creation of new color themes. In contrast, open-ended chat-assisted software restructuring in the visualization showed mixed results. This suggests a need for stronger guardrails for the employed LLM and better incorporation of user feedback. Overall, the assistant was perceived as usable and promising for reducing interaction overhead during exploratory program comprehension tasks. We provide a video presenting the chat assistant's use and a reproduction package of the software system that was used in our evaluation.

0.46

Analysis and Visualization of the Linux Kernel's Software Evolution Using the City Metaphor

cs.SE ·2026-10-07 ·Malte Hansen, Lukas Damerau, Moritz Bugla, David Moreno-Lumbreras, Wilhelm Hasselbring
publishedcode/data
Abstract · 4 pages, 2 figures, accepted at VISSOFT 2026

The Linux kernel is one of the largest and longest-maintained open source projects in existence. With more than 40 million lines of code, understanding the kernel's internal structure and assessing its software evolution is a great challenge. In this paper, we present an approach to visualize the Linux kernel using our software visualization tool ExplorViz. We analyze commits from the Linux Git repository using a custom analysis service. The web-based frontend utilizes the 3D city metaphor for visualization of the software structure. Metrics such as the number of lines are collected for each file and are accumulated for directories and commits. Calculated metrics can be mapped to the building's dimensions or be displayed via a heat map. A wide range of visualization, search, and filter options enable the interactive exploration of the visualized data. We present both a visualization of all files in the kernel repository and a visual analysis of the code evolution of Rust files in the Linux kernel. Demo video: https://youtu.be/3ODnERqD8g0

0.45

Using Small Language Models to Reverse-Engineer Machine Learning Pipelines Structures

cs.SE cs.LG ·2026-10-07 ·Nicolas Lacroix, Frederic Precioso, Mireille Blay-Fornarino, Sebastien Mosser
Abstract

Context: Once defined a taxonomy of stages structuring Machine Learning (ML) pipelines (e.g. Data Preprocessing, Modeling...), extracting these stages from source code is key for better understanding ML practices. However, the diversity caused by the constant evolution of ML (e.g., algorithms, datasets) makes this task challenging. Existing approaches either rely on non-scalable manual labeling or on classifiers that do not properly support domain's diversity. These limitations call for more reliable solutions. Objective: We evaluate whether Small Language Models (SLMs) can leverage their code understanding and classification abilities to address these limitations, and enhance our understanding of practices in ML. Method: We conduct a confirmatory study based on two relevant reference works representing current limitations in the state-of-the-art. We first compare several SLMs using Cochran's Q test, then evaluate the best-performing model against reference studies via two McNemar's tests. An additional Cochran's Q test examines how taxonomy definition variations affect the SLM performance. Finally, goodness-of-fit tests compare ML practice insights from SLM classification with those from prior studies. Results: First, we found that the taxonomy wording significantly impacts classification performance. Second, the best performing SLM yielded good results, yet, without outperforming other classifiers. Third, the three classification methods led to significantly different insights, with varying effect sizes, when exploring practices of data scientists. Conclusions: Limitations of existing classification methods bias our understanding of ML practices. While current SLMs show promising results without prior fine-tuning, they still exhibit common limitations, in addition to inference high costs challenging their applicability in large-scale studies.

0.45

Designing Collaborative AI-Driven Workflows for Scientific Software Engineering

cs.SE ·2026-10-07 ·Akash Dhruv, Max Knobbe, Pedro Machado, Anshu Dubey
Abstract

Agentic artificial intelligence systems can carry out a broad range of tasks in software engineering and scientific research, from writing and translating code to running workflows for data analysis and visualization. In scientific computing, the difficulty is verifying that agent-generated code is both correct and understandable to teams whose members bring different areas of expertise. We therefore argue that these systems are best used within collaborative team structures rather than as full automation. In the workflows we propose, domain experts write the specification and plan, and agents operate under a deterministic orchestration pattern to write the target code. Each stage ends with a numerical comparison against the reference code and requires human review and approval before the next begins. We evaluate these workflows on the translation of a large high-energy physics application from Fortran to C++, running the same task under different orchestrators, design patterns, and models. Across fourteen experiments, a simple author--reviewer loop with enforced limits completed a comparable number of files to a multi-agent workflow at about one-third of the cost per file.

0.45

Code Understanding is a Bottleneck for Coding Agents

cs.SE cs.AIcs.PL ·2026-10-07 ·Nishant Balepur, Kiran Tomlinson, Tobias Schnabel
position
Abstract · In-Progress Preprint

Repository benchmarks (e.g., SWE-bench) for coding agents often assume that lines of code edited can predict task difficulty, but such datasets' poor control over code and task types makes it hard to know which abilities truly drive agent errors. We present CABRA: a Coding Ability Blueprint for Rigorous Agent evaluation. CABRA builds tasks from scratch as call graph transformations and scales difficulty via a task size parameter on four axes: function traversal, search, runtime resolution, and instruction following. We run eight LLMs and six coding agents on 6,840 CABRA tasks to show: 1) LLM accuracy falls as task size~grows, but agents stay near-perfect by offloading work to tools (e.g., grep); 2) Larger CABRA tasks elicit more tool calls for reading and analysis, while a separate study on SWE-bench Verified shows these tool call counts predict agents' accuracy better than lines of code edited, suggesting task difficulty for agents can lie in understanding code to edit, not just in making edits; 3) Extending CABRA to an intense understanding task where models analyze divergent logic across two classes backs this finding, as agent accuracy finally falls. More broadly, we argue for synthetic evaluations like CABRA to unmask LLM weaknesses trivialized by tools (e.g., needle-in-a-haystack) and abilities beyond just editing (e.g., understanding) that coding agents still find difficult, pairing SWE-bench-style tasks with controlled diagnosis.

0.45

Probabilistic Sensing, Deterministic Authority: Admitting Model-Produced Observations into Sufficiency-Checked Governance Contracts

cs.SE cs.AI ·2026-10-07 ·Gaston Besanson
code/data
Abstract · 23 pages. Sixth paper of the SARC series. Preregistered; external reviews and an independent reproduction are committed in the repository. C…

When a field that an authority contract needs exists only in unstructured evidence, a model can sense it. We admit the model's output only as an observation record with a score. An admission policy, with thresholds fitted on a held-out split at a declared false-positive ceiling, maps each score to true, false or unknown. Unknown denies. A deterministic, sufficiency-checked contract decides. The probability that sensing changes the verdict is bounded by the sum, over the contract's sensed fields, of the admitted-wrong and unknown rates. This is an instantiation of union-bound reasoning, indexed by the contract. Minimising the estimated bound is a valid cost model for choosing among sufficient contracts. In a registered study on two constructed domains with two sensor families (36,000 model calls), no cell refuted the bound. Deny-to-allow changes from sensing appeared for the first time in this programme: 13 of 21,000 test verdicts, all from 3 contradictory records; each flip in a cell with a registered bound lay under it. Sensing-aware selection picked the lower-exposure contract in 4 of 4 registered tests. Both sensors' scores were informative but not calibrated. Correctness is relative to the declared loss model, candidate representation and reachable states; all domains are constructed.

0.44

TaoD2C-Bench: Benchmarking MLLMs for Industrial UI Code Generation Beyond Visual Fidelity

cs.SE cs.AI ·2026-10-07 ·Chengwei Shi, Yunnong Chen, Tingting Zhou, Qiang Lu, Shiyu Yue +3 more
benchmark
Abstract · 26 pages

A key challenge for multimodal large language models (MLLMs) is moving beyond visual recognition to constraint-aware cross-modal reasoning. This involves combining visual cues with information from other modalities to understand elements' relationships under domain-specific rules. This challenge is acutely evident in industrial design-to-code (D2C), which converts user interface (UI) designs into code and requires MLLMs to connect design images with disorganized layer metadata, infer component and layout implementation requirements, and realize them in code under target-library constraints. However, these capabilities remain insufficiently evaluated in realistic industrial settings. To fill this gap, we present TaoD2C-Bench, a benchmark for evaluating MLLMs' ability to generate UI code that satisfies implementation requirements in industrial applications. The TaoD2C dataset consists of 2,861 production designs from 17 commercial platforms with 97,652 expert annotations across four categories: Component, Group, Alignment, and Position. These annotations distinguish required constraints from permitted implementation choices. TaoD2C-Bench defines three tasks: end-to-end UI code generation, requirement inference, and requirement realization. Evaluating eight MLLMs reveals substantial gaps in generating UI code that satisfies implementation requirements, alongside distinct performance profiles in inference and realization. We further show that MLLMs' visual reconstruction ability does not necessarily imply an ability to generate code that meets these requirements. We release TaoD2C to support research on industrial UI code generation.

0.39

TestJack: Should you trust the results in coding benchmarks? Agentic Coding Benchmarks Auditing via Evaluator Evolution

cs.SE cs.AI ·2026-10-07 ·Shuangjie Yao, Hao Wang, Koushik Sen, Simin Chen, Baishakhi Ray +1 more
benchmark
Abstract

Large language model (LLM) agents are rapidly reshaping software engineering, accompanied by an explosion of new code benchmarks. Yet nearly all existing benchmarks still rely on the same decades-old criterion: a solution is correct if it passes a fixed set of unit tests. Such tests are often insufficient: they check only part of what the task requires, so agents can reward hack them or silently miss required behavior while still passing every test. As a result, higher benchmark scores may partly reflect better adaptation to the evaluator rather than better problem solving. Existing works focus on static test augmentation: they strengthen each task's tests once, before any trial is seen, and thus overlook how real trials actually fail. We introduce TestJack, a scalable framework for evaluating patches beyond fixed tests. For each trial, TestJack generates tests targeting prompt requirements the patch may violate, retains only tests passed by the ground-truth patch, and re-examines any trial failures. Each confirmed failure is thus supported by a replayable test. To reduce evaluation cost, we also introduce a lightweight variant which audits a random sample of trials in depth and reuses the resulting tests across all trials for the same task. Across 6 frontier model backends and 5 benchmarks such as DeepSWE and SWE Marathon, we find that about 34.4% of the model trials currently judged correct violate the task requirements, lowering the overall resolution rate from 50.6% to 33.2%. Our results reveal a fundamental limitation of current coding-agent evaluation: as LLMs become better at optimizing against fixed evaluators, those evaluators themselves must become more adaptive.

0.39

Cross-Provider Review as a Runtime Contract for Coding Agents: A Controlled Pilot and Fault-Injection Study

cs.SE cs.AI ·2026-10-07 ·Bowen Xu, Boyu Chen
code/data
Abstract · 11 pages, 1 figure, 3 tables. Project page: https://openterminalagent.surge.sh/

Coding agents increasingly share a workstation while drawing on separate providers and subscription allowances. A second agent can inspect a completed answer, but the call spends another pool and may provide no substantive finding. We describe an advisory cross-provider review contract: distinct resource pools, bounded execution, restricted reviewer capabilities, complete input delivery, usable semantic output, explicit failure states and durable per-attempt evidence. In a controlled, agent-authored pilot of 20 paired development turns, eight had a material reviewer finding (95% exact interval 19.1-63.9%). A boundary-condition scan across both reviewer backends reproduced a previously discovered false success on partial input: four truncation levels passed historically and failed after repair. The scan also found and repaired cancellation during process reaping. In real CLI probes, Claude had no writing tools; Codex attempted writes in five of five read-only trials, each write tool failed, and no disposable repository changed. These tests cover specified paths and versions, not field reliability. A preregistered shadow study of metadata-only review allocation accrued 25 formal observations before an exact-runtime regression found a third defect: a reviewer exiting nonzero with a well-formed verdict was counted as complete. Exit status was not recorded per attempt, so exposure cannot be resolved retrospectively. The 25 formal and two pending records remain an audit cohort; the measurement-valid cohort restarted at zero and collection has begun. No gate result is reported.

0.39

Why Software Engineering Is Indispensable in the Age of Coding Agents

cs.SE cs.AI ·2026-10-07 ·Alfonso Fuggetta
published
Abstract · Accepted for publication in Communications of the ACM. 6 pages. v2: final version submitted to the publisher; Section 1 updated with evidenc…

Can AI make Software Engineering (SE) -- the discipline -- obsolete? And can it make software engineers -- the professionals -- redundant? This paper argues that the rise of capable AI coding agents makes SE and software engineers essential, not obsolete: the missing foundation without which AI-assisted development produces misleadingly plausible, unverifiable, and ultimately untrustworthy software. Three structural properties of large language models (probabilistic generation, agnosticism, and semantic statelessness) create a structural vacuum that no amount of training can eliminate. Filling it requires four knowledge levers: methodological knowledge, domain knowledge, design choices, and process choices. All four must be reified as persistent artifacts, and each requires the software engineer as methodologist, mediator, and custodian.

0.37

An explicit shuffle construction of the homotopy span model of differential linear logic

cs.LO ·2026-10-07 ·Paul-André Melliès, Federico Olimpieri, Lionel Vaux Auclair
published
Abstract · To appear in the Proceedings of MPFS 2026

We give a direct and explicit construction of the homotopy span model of linear logic originally formulated by Melliès using abstract homotopy theory in the Quillen category Cat of small categories equipped with its natural model structure. Following the principles of homotopy theory, the formulas and proofs of linear logic are interpreted in the model as small categories up to categorical equivalence, the notion of weak homotopy equivalence underlying the natural Quillen model structure on Cat. Here, we explain how the original interpretation of proofs as fibrant spans can be replaced by a more liberal interpretation of proofs as separately fibrant spans. This relaxation from fibrant spans to separately fibrant spans enables us to give a simple and intuitive description of the homotopy span model, where the axiom and cut links are interpreted as identity spans, and where the symmetries of the exponential modality are captured by a shuffle structure on the interpretation of proofs.

0.36

A new method for proving confluence on abstract reduction systems --- Confluence of non-E-overlapping weakly-shallow TRSs ---

cs.LO ·2026-10-07 ·Masahiko Sakai, Mizuhito Ogawa, Michio Oyamaguchi
Abstract · 68 pages, 17 figures

This paper proposes a new method for proving the confluence of an abstract reduction system (ARS) by clarifying the sufficient conditions, called compatibility and edge commutativity, for expanding a given finite sub-ARS into a confluent one by adding rewrite edges. This method can be regarded as an extension of our earlier work, which showed that a weakly non-overlapping, shallow, and non-collapsing term rewriting system (TRS) is confluent. Furthermore, we apply our method to demonstrate that a non-$E$-overlapping and weakly shallow TRS is confluent. Here, a term is weakly shallow if each defined function symbol occurs either at the root or in the ground subterms, and a TRS is weakly shallow if both sides of all its rewrite rules are weakly shallow. This drops the non-collapsing condition assumed in our previous work on weakly shallow TRSs. Moreover, since a weakly shallow TRS is non-$E$-overlapping whenever it is non-$ω$-overlapping, and the latter property is decidable, we also obtain a decidable sufficient condition for confluence: non-$ω$-overlapping and weakly shallow TRSs are confluent.

0.34

Applying Security by Design at the Point of Execution: How Governed Security Requirements Affect the Security of AI-Generated Code

cs.SE cs.CR ·2026-10-07 ·Pedro Farinha
benchmarktheorycode/data
Abstract · 17 pages, 3 figures, 4 tables, 7 appendices. Artefact: doi:10.5281/zenodo.23213449 and doi:10.6084/m9.figshare.34168182

Security by design asks that security requirements are defined before code is written. Secure-code benchmarks typically measure the opposite situation: the agent receives the task without requirements and is scored with security tests it has not seen. We measured what changes when security requirements, selected from a governed security-by-design knowledge base (SbD-ToE) and delivered through a Model Context Protocol (MCP) server, are given to the generator at the point of execution. On DualGauge, which scores with a language-model judge, we ran 59 Python tasks: the share of tasks passing all security tests rose from 44.1% to 78.0% and the share of security tests passed from 77.4% to 93.0%. On BaxBench, which runs functional tests and real exploits in containers, we ran 28 backend scenarios: among functionally correct solutions, the share with no successful exploit rose from 65% to 86%, comparable to the authors' Oracle Security Reminder (85%), which they call an unrealistic upper bound; the requirement delivery reached it without knowing the tests. The study has limits. The pre-specified joint secure-and-functional metric did not improve significantly on either benchmark. The code written with the governed requirements is stricter, and it failed functional tests that fix values the task never states; we attribute those failures case by case in a post-hoc analysis. Each experiment used one model and one generation per task. The requirements come from the author's open-source manual; we do not assess its completeness. Governed security requirements delivered at the point of execution appear to improve the security of generated code. Current benchmarks cannot tell whether that code meets the security requirements that apply to it, only whether it shows the weaknesses their evaluators were built to detect. We propose measuring conformance with the applicable requirements.

0.34

TestGRAD: Evolving Test Suites via Failure Pattern Momentum for SWE-Agent Ensemble

cs.SE cs.AI ·2026-10-07 ·Pengfei He, Jiayuan Zhou, Shaowei Wang, Ruiqi Pan
Abstract · 12 pages, including 3 pages of supplementary material

SWE-agent ensembles improve issue resolution by combining candidate patches from different agents with complementary strengths. The central problem is therefore test-based selection: generate tests, execute candidate patches, and identify the best patch. We formulate this process as test-space optimization: evolving an executable repository test suite until it distinguishes competing patches. Existing test-generation methods are limited optimizers. They usually lack an explicit loss for ensemble selection, optimize through incomplete directions that mostly create new tests or delete old ones, and perform one-off generation without feedback from repeated failures. Inspired by gradient descent with momentum, we introduce TestGRAD, a framework for automatic test optimization. TestGRAD centers on three concepts. Differential loss gives the optimizer an explicit execution-defined target: useful tests should separate candidate patches by behavior. Full CRUD gradients expand the update direction from merely creating or deleting tests to reading existing test infrastructure, creating new tests, updating stale assertions, and deleting only obsolete tests. Failure Pattern Momentum mines frequent failure sequences from memory, allowing the optimizer to avoid repeated non-discriminative directions while compressing the failure-history context. On SWE-bench Verified, TestGRAD achieves 84.2% Pass@1 with a 4-agent ensemble, outperforming the strongest baseline (80.6%) by an absolute improvement of 3.6 percentage points, while compressing failure-history context by over $100\times$.

0.31

Modal Extensions of Generalised Nelson Logics

cs.LO ·2026-10-07 ·Igor Sedlár, Pietro Vigiani
Abstract

Motivated by relevant epistemic logic, we extend Dunn's simple Kripke-style binary relational semantics for the semi-relevant logic RM to fit a modal extension of RM. It is shown that the modal axioms and rules that need to be added to RM to obtain a sound and complete axiomatisation of its modal extension give an axiomatisation of modal extensions of all logics in the family of generalised Nelson logics, also studied by Dunn.

0.31

Ruleless Digital Twins: Toward Declarative Decision-Making Through Standardized Frameworks and Technologies

cs.SE ·2026-10-07 ·Ivan Spajić, Volker Stolz
Abstract · 47 pages (bibliography included), expanded version of the Spajić and Stolz (2026) DataMod paper

Digital twins (DTs) can be thought of as digital counterparts of physical objects, or more generally, twinning targets (TTs). To enact changes on and optimize for properties of their TTs, DTs use an element of decision-making. Traditionally, many domains, such as home automation, utilize rule-based decision-making models to imperatively define DT actions based on desired TT conditions. With evolving user specifications, rule-based models have become increasingly complex to develop and maintain, particularly in scenarios involving optimizations under dynamically changing systems, such as those affected by weather or dynamic energy pricing. We present an alternative in the form of ruleless digital twins (RDTs) that automatically produce optimal decisions with respect to purely declarative user specifications, similarly to the well-established ruleless approach of model predictive control. They do so through a combination of a semantic knowledge base, logical inference, simulation models, and an autonomic computing architecture, each exclusively based on a widely-used standard or technology. We evaluate our proof of concept through an incubator case study and experiments against a rule-based bang-bang controller in the context of a virtual office room environment. Results show core that RDT functionality performs better than the control in terms of maintaining desired room temperatures and minimizing energy costs under dynamically changing spot pricing. Additionally, due to combinatorial decision tree construction, RDTs require considerable computational power for longer prediction horizons, although their applicability to specific problem domains (and use of shorter horizons) ultimately falls on their users.

0.31

A repeating sequence of simple rules creates a universal computer

cs.FL nlin.CGcond-mat.stat-mechnlin.AO ·2026-10-07 ·Riz Fernando Noronha
unusual
Abstract · 6 pages, 4 figures

We show that the Turing complete rule 110 in elementary cellular automata can be emulated by alternating three simple symmetric rules in space. As the constituent rules themselves are trivial, this shows that diversity alone can generate emergent phenomena. The rules cannot distinguish left from right, and the chirality of rule 110 is created by their arrangement. Random initial conditions lock on to rule 110 over 90% of the time, explained through the motion of domain walls between desynchronized regions. Adding random noise creates errors that seed domain walls, whose fraction scales with the square root of noise strength, explained through a simple rate equation. This points to a route to complexity through repeated, perturbed units, common in biology, and shows that in cellular automaton models of quantum field theory, broken symmetries such as parity need not be present in the local rules.

0.28

Design-Time Conformance Checking for Pulse-Level Quantum Control

cs.SE quant-ph ·2026-10-07 ·Rylan Malarchick
code/data
Abstract · 6 pages, 1 figure, 3 tables. Code: https://github.com/rylanmalarchick/qconform (release v0.1.0)

A pulse-level quantum control program is written against a device whose limits are recorded, if at all, in vendor documentation and source code. A program that exceeds them can be refused at compile time. It can also be accepted and silently altered, or compile and then fail at the board. In the last two cases the experiment runs, and the data does not correspond to the program that was written. We present qconform, a checker that decides whether a pulse program is realizable on a device, given a versioned capability descriptor for that device. Every constraint in a descriptor cites the toolchain observation that established it. The checker is design-time, offline, and deterministic, and it uses no floating point. It reports a verdict, the rules it applied, and a coverage manifest that names what it did not check. We evaluate qconform by differential testing against the QICK and Qblox toolchains, on three QICK board configurations and one Qblox cluster. Over 1263 program runs of 969 distinct programs, the checker accepted no program that a toolchain refuses. On both vendors, a descriptor pinned to one toolchain release accepts programs that older releases refuse. The evaluation also found ten defects in the checker and its descriptors, all fixed.

0.27

Beyond Type-checking: Towards Holistic Evaluation of Formal Specification Generation

cs.SE cs.AI ·2026-10-07 ·Srijith Nair, Aditya Vempaty, Jia Liu, Ashish Jagmohan
publishedtheory
Abstract · Accepted at NeurIPS 2026 Workshop on AI for Verifiable Coding (20 pages, 7 figures)

When generating verifiable code, natural language requirements are mapped to machine checked code using LLMs and agentic workflows. A crucial component of this pipeline is specification generation (SpecGen), which produces a formal contract against which an agent can prove implementation correctness. Proof generation can obtain deterministic feedback from a theorem prover, but SpecGen lacks a definitive check that a generated specification captures the user's intent. A checked proof can therefore establish correctness against a specification that misrepresents the intended behaviour. We take a step towards holistic SpecGen evaluation with a unified dataset assembled from $350$ existing Lean tasks, including $189$ from VERINA and $161$ from CLEVER, and a framework covering formal validity, reference similarity and equivalence, and behavioural adequacy. We distinguish acceptance of required inputs from acceptance of valid outputs and rejection of invalid outputs, while making each metric's evidence scope explicit. Across four SpecGen configurations, restricting the generalized tree edit distance (GTED) comparison, a reference similarity measure, to $32$ jointly measurable VERINA tasks changes the VERINA configuration's position from second to fourth in mean similarity, showing the importance of measurement coverage. In an authored control, a specification achieves $100\%$ positive test recall and negative test rejection while accepting $0\%$ of required inputs. This demonstrates that perfect postcondition scores can miss an unusable input contract, motivating separate feedback on input coverage and output constraints.

0.27

Can LLMs Simulate Novice Programmers' Misconceptions?

cs.SE cs.PL ·2026-10-07 ·Batuhan Yeltekin, Daniel Bauer
Abstract

We evaluate 13 LLMs on generating, solving, simulating, and diagnosing novice programming misconceptions on code-tracing problems. While frontier models reliably perform these tasks, small models ($\le$14B) struggle with dynamic execution. Notably, code-tuned models fail during misconception simulation by reverting to correct execution. Finally, misconception diagnosis is significantly easier in True/False formats than in open-ended generation tasks.

0.26

AdaT$^2$: Adaptive Test Transformations for Black-Box Boundary Testing of Conversational Agents

cs.SE ·2026-10-07 ·Liting Lin, Boxi Yu, Qinghua Xu, Yuzhong Zhang, Lionel Briand +1 more
benchmark
Abstract

Conversational agents based on large language models (LLMs) must comply with policies. Each condition in a policy draws a boundary between user requests, and the agent must behave differently on its two sides. We present AdaT$^2$, which extracts statements from the agent's replies in exploratory conversations with an LLM acting as the user, and uses the statements to guide boundary test generation. Each statement describes one condition and the behavior expected when the condition holds. Besides plain tests guided by single statements, AdaT$^2$ writes transformed tests guided by pairs of a statement and a test transformation instruction, such as "omit one required input". The instruction of a pair can move a test to the other side of the statement's boundary or to another boundary. The statements and instructions form far more pairs than a run can try, and many pairs are not applicable. Adaptive pair selection therefore chooses the statement of each pair by novelty and the instruction with the bandit algorithm Bayes-UCB, which learns from whether earlier pairs yielded a test and whether the agent passed it according to an LLM judge. Our benchmark counts the two sides of each boundary separately and distinguishes boundaries explicitly defined by the agent's prompt, tool code, or knowledge base from boundaries that the agent's LLM infers from domain knowledge. On four domains of $τ^3$-bench, 62.7% to 83.3% of AdaT$^2$'s tests are valid boundary tests whose expected behavior is explicitly defined, higher in every domain than AgentEval's (47.7% to 68.1%). Transformed tests add 13 to 46 explicitly defined boundaries that plain tests miss. As regression tests, AdaT$^2$'s test suites detect all eight seeded policy faults in the airline domain and four of eight in the retail domain, and AgentEval's test suites, with fewer than a third as many tests, detect five and two, respectively.

0.24

CTML: A Comprehensive Framework for Continuous Testing in Machine Learning Operations

cs.SE ·2026-10-07 ·Bestoun S. Ahmed
Abstract

The increasing use of machine learning (ML) systems in production has highlighted some important weaknesses in traditional software testing methods. Unlike regular software, ML systems can fail in specific ways due to their reliance on data, including issues like data drift, concept drift, and performance decline. To address these issues, this paper introduces the Continuous Testing for Machine Learning (CTML) framework. This framework is designed specifically for MLOps environments and offers a thorough, automated solution. It establishes a standardized way to evaluate ML systems across four main areas: data quality, drift detection, model quality, and pipeline validation. The CTML framework consists of a modular setup that includes a Test Registry, Test Runner, Result Store, and Integration Layer. This design ensures it integrates smoothly with existing Continuous Integration/Continuous Delivery (CI/CD) pipelines and MLOps platforms. We put this framework to the test in a predictive maintenance case study, which showed how effective it is at identifying issues like data drift, performance degradation, and bottlenecks in the pipeline problems that traditional testing methods might overlook. The findings indicate that CTML successfully uncovers critical failures in the system across all testing dimensions, providing useful insights to help keep ML systems running smoothly in production. Overall, this work presents a new, integrated framework that connects academic research with industry best practices, pushing forward the standards in MLOps quality assurance.

0.24

When Sub-Agents Work in Parallel: The Promises and Pitfalls of Dynamic Concurrency in Long-Horizon Coding Tasks

cs.SE ·2026-10-07 ·Han Li, HanHaoNing Li, Ziqian Jiang, Yiling Lou
Abstract

As coding agents advance from bounded software engineering tasks toward long horizon development, dynamic concurrency offers a promising way to scale complex development tasks. Under this policy, agents decide during execution whether and how to spawn concurrent sub-agents. Model capability largely determines outcomes on shorter tasks, whereas long horizon development makes orchestration central to task completion. Existing work, focused on coding agent failures on shorter tasks or collaboration in predefined multiagent workflows, offers little insight into dynamic concurrency in frontier agents across task complexity. We study dynamic concurrency as an execution policy through controlled comparisons of matched Codex, Claude Code, and Kimi Code executions with the policy enabled or disabled. Across 354 tasks and 2,124 executions spanning a range of task complexities and execution horizons, we evaluate its end to end effects and scheduling behavior, and analyze matched trajectories to characterize 13 concurrency specific failure modes, 28 observable patterns, and the conditions under which it provides an advantage.

0.22

Beyond FAIR: A Fitness Function Framework for Sustainable Research Software

cs.SE cs.CR ·2026-10-07 ·Irdin Pekaric, Philipp Zech
Abstract

Research software sustainability is often assessed through the FAIR principles: findability, accessibility, interoperability and reusability. While important, FAIR does not cover all relevant sustainability concerns. Research software should also be environmentally responsible and secure over time. Excessive resource consumption increases environmental cost, while insecure software creates maintenance overhead, technical debt and barriers to reuse. To this end, in this paper, we extend a previously proposed fitness function framework for sustainable research software beyond FAIR. We introduce two additional sets of fitness functions: environmental functions targeting resource efficiency and execution footprint, and security functions targeting dependency health, vulnerability exposure and secure configuration. Together, these functions broaden continuous software assessment from FAIR compliance to a more complete view of sustainability. The resulting framework treats research software sustainability as a multidimensional property that includes FAIR, environmental responsibility and long-term security.

0.18

GRAML: Graph-Grounded Reasoning and Multi-Task Learning for LLM-Based Software Vulnerability Detection

cs.SE ·2026-10-07 ·Ruohan Li, Miaoqing Tian, Haipeng Qu, Ruobing Jiang
Abstract

Large Language Models (LLMs) have been widely applied to software vulnerability detection. However, their performance is often limited by insufficient use of control-flow and data-flow information. In this paper, we propose GRAML, a framework that combines graph evidence, vulnerability description generation, and multi-task training. GRAML first performs static analysis on C/C++ programs to extract critical source lines and typed line relations as structural evidence. It then uses this evidence to guide GPT-5 through the Tree-of-Thought-guided Vulnerability Reasoning (ToT-VR) process and generate vulnerability descriptions. These descriptions are further combined with Detection, Localization, and Assessment samples to build a unified four-task training dataset. We evaluate GRAML on an in-distribution (ID) test set and six out-of-distribution (OOD) datasets. The results show that GRAML achieves average F1 scores ranging from 66.67% to 68.70%, outperforming state-of-the-art baselines by up to 30.92%. Ablation experiments further show that ToT-VR and graph-guided vulnerability descriptions improve detection performance compared with standard Chain-of-Thought (CoT) reasoning and raw Code Property Graph (CPG) serializations. These findings provide practical guidance for building more reliable and secure software engineering systems with large language models.

0.16

Bayes Rule for Conditioning Probabilistic Programs

cs.PL cs.LO ·2026-10-07 ·Feras A. Saad
Abstract

Probabilistic programs provide a rigorous programming-language foundation for expressing rich probabilistic models and offer language-level support for solving challenging statistical inference problems. A key theoretical challenge that probabilistic programs inherit from probability theory is conditioning a probability distribution on an observation that has probability zero. Many languages address this challenge by using computational rules for conditioning that may work well for some problems but are statistically unsound for others. This article introduces a principled mathematical approach for conditioning probabilistic programs. We first show that existing techniques based on "soft conditioning" and "neighborhood limits" exhibit semantic issues from the perspective of measure theory and produce inference results that disagree with sound statistical reasoning. By leveraging tools from geometric measure theory, we construct a novel disintegration-based semantics for probabilistic program traces that enables sound conditioning. The theory delivers an explicit Bayes rule representation of the conditional distribution, which is directly compatible with widely used probabilistic inference algorithms that require access to unnormalized posterior densities. Our Bayes rule handles probabilistic models with loops, branches, discrete, continuous, and mixed-type distributions, observations on manifolds, differentiable structure, and additional modeling motifs that arise in practice.

0.15

Has LLM Screening Performance Stalled in Software Engineering Systematic Reviews?

cs.SE cs.AI ·2026-10-07 ·Aleksi Huotala, Miikka Kuutila, Mika Mäntylä
benchmark
Abstract · 53 pages, four external figures available in the research artifact

Screening in systematic reviews (SRs) is manual and time-consuming. Prior work has explored large language models (LLMs) for automating this step, but LLMs are evolving rapidly, so earlier performance claims may no longer accurately reflect their screening performance. We used an existing software engineering SR screening benchmark (SESR-Eval) as our data. We also power-sampled a new, smaller dataset (SESR-Eval-Mini) that allows evaluation at lower costs. Using this data, we evaluated eight new LLMs for screening performance. Additionally, we tested different prompts, analyzed LLM agreement in screening decisions and criteria, and examined the effect of refining the inclusion and exclusion criteria on screening performance. The eight new LLMs performed marginally better than the seven old ones: avg. MCC across secondary studies rose from 0.347 to 0.365. Differences between secondary studies are still bigger than between LLMs. Computing the overall screening decision from criterion-level decisions degraded screening performance only slightly. LLMs generally agree with each other in their corresponding screening decisions (mean Gwet's AC1 = 0.830), though certain inclusion and exclusion criteria showed larger disagreement than others. Refining the inclusion and exclusion criteria slightly improved recall and made decisions easier for some LLMs, but overall impacts of criteria refinement were modest. LLMs are not yet ready to replace humans in paper screening and the advantages new, more costly models bring, appear to be very limited. Agent-based approaches, prompt engineering, and further criteria refinement are three potential future research avenues.

0.13

AgentTracer: Tracing Indirect Prompt Injection Attack through Fine-Grained Intention-Execution Alignment

cs.SE ·2026-10-07 ·Zitong Yao, Jiangrong Wu, Yixi Lin, Anrui Huang, Luoyun Zhang +1 more
Abstract

Large language model (LLM) agents interact with external resources to complete complex user tasks, exposing them to indirect prompt injection (IPI), where malicious instructions redirect agents toward attacker-intended tasks. Since IPI is difficult to defend against in real-world environments, post-incident tracing is essential for locating the injection source and reconstructing the attack chain. However, existing tracing methods primarily capture explicit control-flow and data-flow dependencies, overlooking the implicit relationships among tool calls driven by the malicious instruction. These tool calls may lack explicit dependencies and be interleaved with legitimate operations, making complete attack-chain reconstruction difficult. In this paper, we present AgentTracer, an intent-aware tracing framework that treats IPI as task intent drift. AgentTracer recovers implicit decision dependencies among tool calls to construct an Intent-Driven Execution Graph that connects dispersed tool calls by task intent. It combines the user request with an operation knowledge base to construct a user intent authorization space and identify intent-drift tool calls. Starting from an anomalous tool call under audit, AgentTracer performs target-based pruning and backward tracing to reconstruct the attack chain and locate the injection source and injection point. To evaluate AgentTracer in the presence of noise from normal tasks, we combine execution logs constructed from successful IPI attacks in AgentDyn and InjecAgent with normal execution logs containing 1,800 user requests and 9,000 background tool calls without IPI. In end-to-end experiments, AgentTracer achieves 94.17 percent injection-point accuracy and 93.56 percent path precision. Comparative experiments show that AgentTracer improves injection-point accuracy over existing methods by 18 to 54 percent.

0.11

QuSema: Detecting Silent Bugs in Quantum Libraries via Quantum-knowledge-enhanced Agents

cs.SE cs.AIquant-ph ·2026-10-07 ·Yujin Song, Kaining Zhang, Qixin Zhang, Shuai Wang, Pingchuan Ma +1 more
benchmark
Abstract

Quantum libraries are now critical infrastructure for quantum algorithm development, yet their correctness remains difficult to test. Existing testing techniques mainly rely on failure-based or comparison-based oracles, exposing bugs only when executions fail, violate runtime checks, or disagree with another implementation. Their applicability is limited when suitable execution-based oracles are unavailable, leaving some silent bugs undetected. Such missed bugs can produce incorrect results that propagate into experimental conclusions, simulation studies, and algorithmic designs. Here we present QuSema, an autonomous testing agent for finding silent bugs in quantum libraries. QuSema uses constraints from quantum semantics and documentation as a source-level semantic oracle to assess whether implementation logic can produce invalid outputs from valid inputs. It operates through an agentic loop that repeatedly inspects library API documentation and source code, reasons about the intended behavior of quantum operations, identifies potential semantic deviations, and validates them by generating executable tests through library APIs. Guided by quantum-domain reasoning, QuSema turns high-level behavioral mismatches into concrete, user-triggerable bug reports, enabling it to uncover non-crash defects. We implement QuSema for Qiskit and PennyLane. On a benchmark of 20 historical silent bugs, QuSema achieves higher mean bug relocation counts than Claude Code and Codex, with the DeepSeek configuration costing less than Claude Code. QuSema also discovers 40 previously unknown bugs confirmed by the developers, including 30 silent bugs.

0.04

Theory & Algorithms 79

3SUM Is Really Hard: A Real-to-Integer Reduction

cs.DS ·2026-10-07 ·Nick Fischer, Adam Polak, Jonas Schmidt
top venuetheory
Abstract · To appear in FOCS 2026

We show that if the 3SUM problem on integer-valued inputs can be solved in truly subquadratic time, then it can also be solved in truly subquadratic time on real-valued inputs. This answers an open problem posed by Chan, Vassilevska Williams, and Xu [STOC 2022], and constitutes the first such tight real-to-integer self-reduction in fine-grained complexity. Our proof relies on a surprising combination of the Frank--Tardos theorem from optimization with Freiman-type theorems from additive combinatorics.

0.71

Counting Polyominoes, Revisited: Corrigendum and Addendum

cs.DM ·2026-10-07 ·Gill Barequet, Gil Ben-Shachar
publishedunusualhighly cited
Abstract

This is a Corrigendum and Addendum to Algorithmica paper (2026) 88:51.

0.65

Network Coding Can Beat Routing in Undirected Multiple-Unicast Networks

cs.IT ·2026-10-07 ·Xindan Zhang, Baochun Li, Zongpeng Li
theoryhighly cited
Abstract · 17 pages

Network coding lets the nodes of a network combine messages rather than merely forward them. In undirected networks, where the two directions of an edge share its capacity, Li and Li conjectured in 2004 that coding offers no advantage over fractional routing; the conjecture has been confirmed for many special classes but never settled in general. In this paper, we disprove it by constructing a finite connected simple undirected network with unit shared edge capacities, maximum degree three, and distinct leaf terminals, on which a binary linear block code achieves a common rate strictly above the maximum fractional routing rate. The construction turns a short completion-time code into a reversible circuit whose registers all carry independent messages, and a temporal metric on its wires yields the strict routing bound. The underlying integer-coefficient construction works over every finite field and every nontrivial finite abelian group; for deterministic fixed-schedule codes the common-rate supremum is one and the closure of the rate region is the unit cube. Amplifying the gap with a degree-preserving tensor construction and an even subdivision gives, for every $0<β<1$, an unbounded family of connected, simple, subcubic, bipartite networks with girth at least $n^β$ on which coding approaches rate one while routing is bounded by $K_β/(\log n)^c$, with one positive exponent $c$ independent of $β$. The finite counterexample and the family theorem are formalized in Lean.

0.65

Taxonomic Classification with Tag Arrays

cs.DS q-bio.GN ·2026-10-07 ·Travis Gagie, Gonzalo Navarro
code/dataunusual
Abstract · 9 pages, 1 table. Rewritten: KATKA now finds long MEMs with Boyer-Moore-Li on Movi 2 or an RLCSA, counts genera with run-length, grammar-com…

We present a tool, KATKA, for taxonomic classification with tag arrays. Compared to Ahmed et al.'s state-of-the-art tool Cliffy ({\it Genome Research}, 2025) on the SILVA dataset from their paper and a readset similar to one of theirs, KATKA is faster, more accurate, much smaller and drastically easier to build.

0.64

Minimizing Cumulative Envy in Allocating a Sequence of Items

cs.GT cs.MA ·2026-10-07 ·Paul W. Goldberg, Isaac Robinson, Nicholas Teh
publishedhighly cited
Abstract · 40 pages. Preliminary version of this paper was presented at the 19th SAGT (Sept. 2026)

We study temporal fair division with indivisible goods that arrive sequentially and must be allocated irrevocably. In contrast to the usual online model, we assume that valuations and future arrivals are known in advance, and ask how unfairness evolves during the process. We introduce \emph{cumulative maximum envy}: the sum, over all rounds, of the maximum pairwise envy at that round. Equivalently, this is the area under the worst-envy curve, and it captures both the magnitude and the duration of envy. For a fixed arrival order, we show that the corresponding decision problem is strongly NP-complete and that minimizing this objective admits no constant-factor approximation unless P = NP, even under identical valuations and even under binary valuations. We complement these hardness results with a dynamic program that gives pseudopolynomial-time solvability for a constant number of agents, polynomial-time algorithms in further restricted settings, and an FPTAS for fixed $n$ under identical integer valuations. We then study a sequencing variant where the algorithm may choose the arrival order. This variant remains NP-complete even for two agents with identical valuations; however, a simple greedy algorithm achieves a $3/2$-approximation for $n=2$ agents, an $n/(n-1)$-approximation for any number of agents, and an additive guarantee depending on the maximum value of any good.

0.59

Color Coding for the Sherrington-Kirkpatrick Model

cs.DS math-ph ·2026-10-07 ·Alina Harbuzova, Saba Lepsveridze, Mahbod Majid, Ankur Moitra
Abstract · 33 pages

We give a polynomial-time algorithm for approximating the partition function of mean-field mixed $p$-spin models to arbitrarily high accuracy throughout the second-moment regime. This improves the quasipolynomial-time algorithm of Bencs, Huang, Lee, Liu, and Regts (arXiv:2507.15616) and answers their open question. In particular, our result covers the entire replica-symmetric regime of the Sherrington--Kirkpatrick model. The algorithm is deterministic, runs in time polynomial in $n$ and $1/\varepsilon$, and succeeds for every typical realization of the disorder. Our main technical contribution is a new algorithmic application of color coding to the classical high-temperature expansion introduced by Aizenman, Lebowitz, and Ruelle in their study of fluctuations of the Sherrington--Kirkpatrick partition function. We combine this approach with the zero-freeness established by Bencs et al. (arXiv:2507.15616) to convert additive approximations into multiplicative ones.

0.54

Rectangular matrix multiplication from polynomial products

cs.DS ·2026-10-07 ·Przemyslaw Uznanski
theoryhighly cited
Abstract

In this note, we extend the matrix-multiplication analysis by OpenAI to rectangular products and prove $ω(1,k,1)\le2$ for $0\le k\le\frac{1}{2}$, and $ω(1,k,1)\le1+k+\frac{1}{4k}$ for $k\ge\frac{1}{2}$. In particular, $ω(1, \frac{1}{2},1)=2$ and the dual exponent satisfies $α\ge\frac{1}{2}$. The proof uses explicit degenerations of convolution tensors, multinomial counting, and Strassen's spectral theorem. An angular constraint on three dot-product parameters gives both the rectangular curve and a bound for $ω(a,b,c)$. As an application, Zwick's algorithm for all-pairs shortest paths in directed unweighted graphs runs in $O(n^{2.5})$ time. Combining the rectangular bound with the $(\min,+)$-product improvement of Alman and Vassilevska Williams further gives $O(n^{2.4999})$ running time.

0.54

Smallest String Attractors and Minimal Coverage Certificates of Thue--Morse Words

cs.DS math.COcs.FL ·2026-10-07 ·Simone Faro, Francesco Pio Marino, Arianna Pavone
theory
Abstract

String attractors provide a compact way of representing the complete factor structure of a word: a set of positions is an attractor if every distinct factor has at least one occurrence crossing one of the selected positions. Although the minimum attractor size is known for several classical families of words, describing \emph{all} optimal attractors is typically much more difficult, since it requires understanding the geometry of all factor occurrences rather than constructing a single optimal solution. We give a complete description for the finite Thue--Morse words. Earlier work proved that four positions are necessary and sufficient for every order $n\geq 4$, but the collection of all smallest attractors remained unknown. For every $n\geq 6$, writing $h=2^{n-3}$, we prove that the smallest attractors are exactly the two reflected families where the four offsets are chosen independently from $\{0,1\}$. Hence there are exactly $32$ smallest attractors for every $n\geq6$. The initial cases are genuinely exceptional: $t_5$ has $40$ smallest attractors and $t_4$ has $87$. We also study the attractor condition independently of optimality. For every $n\geq5$, we characterize the complete antichain of inclusion-minimal factor coverages of $t_n$. It consists precisely of the coverages of $aa$, $bb$, and the eight minimal unique substrings of every generation $t_m$, $4\leq m\leq n$. Thus there are exactly $8n-22$ canonical constraints, forming an irredundant exact certificate for attractors of arbitrary cardinality. When attention is restricted to four-position sets, this linear-size system collapses to a constant one: it is enough to test the $24$ minimal unique substrings coming from three consecutive generations, and sixteen of these already force the two optimal families. We also show that three generations are necessary within this natural consecutive-generation hierarchy.

0.52

Min-Plus Convolution Lower Bounds via a Higher-Order BSG Theorem

cs.DS ·2026-10-07 ·Nick Fischer, Ce Jin, Yinzhan Xu
top venuetheory
Abstract · Appears at FOCS '26

Min-Plus Convolution is a central problem in fine-grained complexity, and the associated Min-Plus Convolution Hypothesis forms the basis for a wide range of conditional lower bounds for fundamental problems. It is closely connected to the APSP and 3SUM Hypotheses, and in fact implies both, making it a unifying hypothesis for two of the main pillars of the area. In this work we establish several strong results related to Min-Plus Convolution. We design a universe reduction, showing, under a plausible additive combinatorics assumption, that the Min-Plus Convolution Hypothesis is equivalent to the Strong Min-Plus Convolution Hypothesis. We also obtain tight conditional lower bounds for multiple long-standing problems, including Min-Max Convolution and Bounded Monotone Min-Plus Convolution. Our approach is inspired by Fischer's recent equivalence between several variants of APSP [STOC '26], but extending that technique to the arithmetic setting requires overcoming deep obstacles. To this end, we develop a novel additive structure theorem that can be viewed as a higher-order substitute of the Balog-Szemerédi-Gowers (BSG) theorem, allowing us to extract strong additive structure even from weakly structured sets. Building on this structural result, we show that certain structured 3SUM instances (namely, sets with low rank) can be solved in truly subquadratic time. This algorithm forms the main algorithmic ingredient in our reductions. Besides, it generalizes all previously known truly subquadratic-time special cases of 3SUM, and is therefore of independent interest.

0.50

A Constant-Factor Approximation to Multidimensional Consumer Utility

cs.GT ·2026-10-07 ·Kira Goldner, Taylor Lundy, Thodoris Tsilivis
theory
Abstract · 40 pages, including appendices

Motivated by social services where consumers pay with non-transferable ordeals, we study mechanisms that maximize consumer utility for multiple unit-demand buyers and heterogeneous items whose values are drawn independently from known prior distributions. Prior work in utility maximization approximates social welfare and shows that the gap between optimal utility and social welfare is logarithmic. We resolve the question of whether simple mechanisms can guarantee a constant-factor approximation to optimal utility itself. Our mechanisms achieve a $(5.67+\varepsilon)$-approximation for general independent values, improving to $2e/(e-1)<3.164$ when values are i.i.d. Each buyer chooses their favorite option from posted item prices or free item lotteries, and contention resolution determines which buyers' requests are served; this is BIC, ex-post individually rational, and computable in polynomial time. Our main technical contribution is a general upper bound on the optimal ex-ante-constrained utility that separates the contributions captured by posted prices and free lotteries. These results establish a utility counterpart to the theory of simple, approximately revenue-optimal mechanisms.

0.49

Directed Global Minimum Cut in Almost-Linear Time

cs.DS ·2026-10-07 ·Henry Fleischmann, Jason Li, Thatchaphol Saranurak, Benyu Wang
highly cited
Abstract · 11 pages, 4 figures

We give a randomized reduction from global minimum cut in a directed graph with $n$ vertices and $m$ weighted edges to maximum-flow computations on graphs of total size $\widetilde{O}(m)$. For polynomially bounded integral weights, this yields the first almost-linear $m^{1+o(1)}$-time algorithm [vdBCK+23], improving the previous $m^{1+o(1)}\min\{\sqrt{n},n/m^{1/3}\}$ bound [CLN+22]. Previous work reduced finding the minimum cut to constructing a 1-respecting arborescence, a directed spanning tree with exactly one edge crossing a minimum cut [CLN+22]. They then construct such a tree by sampling from a $(1+ε)$-approximate arborescence packing, which currently requires superlinear time. Our algorithm sidesteps constructing the packing entirely and relies on a novel iterative arborescence sampling procedure. We show that, after $O(\log n)$ rounds of sampling, our final arborescence 1-respects the minimum cut with constant probability.

0.48

Second-Order Image Characterization with Distortion and Secrecy

cs.IT ·2026-10-07 ·Cao Daming
theory
Abstract · Draft manuscript

Image characterization relates coding budgets to the probability carried by source cells and their output images. At second order, refining a cell can reveal unavailable information, while a change of reference measure can alter the relevant likelihood fluctuations. We establish image bounds that account for both effects. A finite-block partition theorem bounds the cost of merging component images by the joint conditional information spectrum of their label. It preserves continuous second-order profiles when that cost is negligible on the square-root blocklength scale. For reconstruction constraints, cell-dependent reference measures retain the actual distortion deficits in a single likelihood. A Gaussian evaluation gives matching second-order rates for Heegard--Berger coding with one uninformed decoder in the strict both-active regime, common reconstruction under a joint agreement and distortion criterion, and two-stage refinement with side information absent at the first decoder. A structured binary source illustrates partition stability even when its global density and conditional-divergence deficits are linear in blocklength. Posterior-mass images also sharpen extraction after a public refinement by charging its residual information rather than its alphabet size. These results preserve the joint constraints and the observations available for choosing the local coding experiment.

0.47

Tweaking the constant in the Linear Hadwiger Theorem

cs.DM math.CO ·2026-10-07 ·David R. Wood
theory
Abstract · v1 This paper was first submitted to the arXiv shortly before the announcement of the disproof of Hadwiger's Conjecture. v2 The main constan…

Norin and Steiner recently proved the Linear Hadwiger Conjecture. That is, there is an absolute constant $C$ such that every graph $G$ satisfies $χ(G)\leqslant C\,\text{had}(G)$, where $χ(G)$ is the chromatic number and $\text{had}(G)$ is the Hadwiger number of $G$. They estimate that their proof gives a non-optimised constant $C$ of order $10^{100}$. This paper uses AI assistance to optimise their argument, and shows that every graph $G$ satisfies $χ(G)\,\leqslant\,1{,}118{,}430\,\text{had}(G)$.

0.47

Making Every Number from 1 to N Under a Fixed Cycle of $+$, $\times$, $-$, $÷$

cs.DM math.COmath.NT ·2026-10-07 ·Sean Lesmana, Theodore Tjugiarto
code/dataunusual
Abstract · 25 pages, 2 figures, 9 tables. Verification code: https://github.com/poscle/AMSD-Number-Generator

Start with the number $2$. At each move, combine two numbers already made, but the operations must be used in the fixed repeating order $+,\times,-,÷$. We ask for the fewest moves needed to make every integer from $1$ to $N$. Since $2$ is already one of the numbers we want and each move makes at most one new number, at least $N-1$ moves are needed. We show that $N-1$ moves are also enough for every $N\ge 9$. Conventional induction cannot work, because a division that comes right after a completed interval $\{1,\dots,P\}$ produces numbers already made. Instead we extend a completed interval $\{1,\dots,P\}$ to $\{1,\dots,3P\}$ all at once, which counting shows is the smallest multiplicative extension $P\to kP$ that can work, and then adjust the last few moves to reach every other $N$. For $9\le N\le 33$ we give explicit sequences, found by computer search.

0.43

MacKay-Neal Codes Achieve Capacity under MAP Decoding

cs.IT ·2026-10-07 ·Kenta Kasai
theory
Abstract · 22 pages

Statistical-mechanical analyses predict that MacKay-Neal codes can achieve channel capacity at fixed degrees. We prove this prediction for the uncoupled $(\ell,3,3)$ MN socket ensemble for every fixed integer $\ell\ge4$. For each binary-input memoryless symmetric channel of capacity strictly greater than $3/\ell$, the ensemble-average block error probability under maximum a posteriori (MAP) decoding is $O(\log N/N)$, where $N$ is the transmitted blocklength. The actual transmitted rate converges to $3/\ell$. The result includes both punctured and transmitted variables and does not condition the sparse square matrix on invertibility. The proof establishes an entropy inequality for six bits subject to even parity by an analytic argument that reduces the domain by symmetry, restricts its interior stationary points, where both partial derivatives vanish, to the diagonal, and controls the resulting scalar functions by explicit polynomial bounds. Exact configuration counts then bound the conditional entropy on the binary symmetric channel. A standard comparison by mutual information extends the entropy bound to general symmetric channels; independent output erasures and a minimum-distance estimate for the transmitted code yield the bit and block error bounds.

0.43

Deterministic Distance Selection in $O(n^{4/3})$ Time

cs.CG cs.DS ·2026-10-07 ·Haitao Wang
top venue
Abstract · To appear in SODA 2027

Let $P$ be a set of $n$ points in the plane. Given an integer $K$ with $1\le K\le {n\choose 2}$, the distance selection problem asks for the $K$-th smallest distance among all pairwise distances of the points of $P$. The problem has been studied extensively over the past four decades. Recently, Chan and Zheng (2023) gave a randomized algorithm with $O(n^{4/3})$ expected running time, while the best known deterministic algorithm runs in $O(n^{4/3}\log n)$ time. In this paper, we present a deterministic $O(n^{4/3})$-time algorithm for the problem, matching the best known randomized complexity. We also consider the more general bichromatic version of the problem, where two point sets $A$ and $B$ are given, with $m=|A|$ and $n=|B|$, and the goal is to find the $K$-th smallest distance among the $mn$ distances between points of $A$ and points of $B$. For this problem, the best previously known algorithm runs in $O((m\log n+n\log m+m^{2/3}n^{2/3})\log(m+n))$ time. We present a new deterministic algorithm with running time $O(m\log^2 n+n\log^2 m+m^{2/3}n^{2/3})$.

0.41

Optimal (Parallel) Spooky Pebbling on Binary Trees

cs.CC ·2026-10-07 ·Mingyu Lee, Sanghyun Lee, Kabgyun Jeong
Abstract · 21 pages, 1 figure

Pebble games model computations under a fixed space budget. Spooky pebbling allows quantum memory to be released by measurement, with the resulting phases corrected later. We study two-input computations with binary-tree dependencies and determine the optimal work and parallel depth for complete binary trees. Our key idea is to clean up the tree in blocks, reducing repeated recomputation of intermediate values. For the complete tree $B_h$ with $n=2^h-1$ vertices and every space budget $h+1\le s\le n$, we give an algorithm with asymptotically optimal work $Θ(nh/\log(s+1))$. At the minimum budget $s=h+1$, this improves the $O(n\log n)$ bound of Kornerup, Sadun, and Soloveichik to $Θ(n\log n/\log\log n)$, resolving their time-optimality question. We also construct a parallel schedule with optimal depth \[ Θ\!\left(h+\frac ns \max\!\left\{\frac{h}{\log(s+1)},\,1+\log^*h\right\}\right). \] Our lower bounds hold for every full binary tree. We also show that achieving optimal parallel depth can require asymptotically more work than minimizing work alone. Applying our schedule to the RNS point-addition trees in the public implementation of Chevignard, Fouque, and Schrottenloher reduces their Toffoli/AND gate count by $20.86\%$ for a P-224 instance and $22.66\%$ for a P-256 instance, using the same arithmetic circuits and peak workspace.

0.41

Simulated annealing and Weak Poincaré inequalities with inverse-polynomial accuracy for the Sherrington-Kirkpatrick model

cs.DS math.PR ·2026-10-07 ·Zhe Hou, Jingcheng Liu, Yixiao Yu
Abstract

We study sampling from the Gibbs distribution of the Sherrington-Kirkpatrick (SK) model with Glauber dynamics. For every fixed inverse temperature $0\leqβ<1$ and every fixed $M,D>0$, we give a polynomial-time simulated annealing algorithm whose output distribution is within $n^{-M}$ total-variation distance of the Gibbs distribution, with probability at least $1-n^{-D}$ over the interaction matrix. The algorithm starts from uniform product spins and uses Glauber dynamics along an increasing inverse-temperature schedule. The same approach also yields partition-function estimates with relative error $n^{-M}$. The key ingredient is a quantitative weak Poincaré inequality. Building on the stochastic-localization approach to weak Poincaré inequalities [Davies, Lee, Sandhu, and Shi, arXiv:2607.08160, 2026], we use Gaussian isoperimetry to directly compare exact stochastic localization paths. Our comparison tolerates failures of local Lipschitz continuity of the posterior mean by controlling their accumulated cost. The local Lipschitzness comes from approximating the posterior means by stationary points of the Thouless-Anderson-Palmer (TAP) free energy in locally strongly convex regions located by approximate message passing, a paradigm introduced by algorithmic stochastic localization [El Alaoui, Montanari, and Sellke, 2025; Celentano, 2024]. The approximation errors are roughly characterized by the validity of the TAP gradient equations, but in a different coordinate. To show that the errors rarely accumulate too much, we need a strong concentration control. A direct argument would require solving a variant of an open problem by Talagrand [2010, Research Problem 1.7.9]. We bypass this open problem with an intermediate conditioning step and transfer the control back by establishing the stability of the TAP gradient under deletion of coordinates.

0.40

The Price of Privacy: Randomness Complexity of Graph-Based Multi-Secret Sharing

cs.IT cs.CR ·2026-10-07 ·Piotr Marszalik
code/data
Abstract · 12 pages, 4 figures, including an appendix with proofs. Code: https://github.com/pjmarszalik/price-of-privacy-code

We study the randomness required to share possibly correlated secret bits among parties connected by a graph. A dealer places shares on the edges so that each party can recover its own secret from its incident shares and learn nothing about the others beyond what its own secret reveals. Anilkumar et al. completely determined the minimum randomness required for three binary secrets. We extend this study to four secrets and obtain results for arbitrary numbers of secrets on general graphs. For four parties on a complete graph, we determine the minimum number of random states for every set of permitted secret combinations: the possible values are one, two, three and four. We also characterize when one random bit suffices on an arbitrary graph.

0.39

The Multiple Unicast Conjecture is False

cs.IT cs.CCcs.DMcs.DS ·2026-10-07 ·Mark Braverman, Zhongtian He
theoryhighly cited
Abstract

The undirected multiple-unicast conjecture [LL04] asserts that network coding offers no throughput advantage over multicommodity flow. We refute this conjecture by constructing a deterministic linear network code over $\mathbb{F}_9$ on a 182-vertex bipartite subgraph of the point-line incidence graph of $\mathrm{PG}(2,9)$. The construction supports $157$ independent unicast sessions at common coding rate at least $1$, while every fractional multicommodity flow has common rate at most $147/157$. By the amplification theorem of~[BGS17], this yields a family of undirected multiple-unicast instances with coding gap $Ω((\log n)^\varepsilon)$ for some $\varepsilon>0$. We also introduce a nondeterministic model of network coding based on locally verifiable certificates, which guides our construction and may be of independent interest. Building on the high-girth graph and error-correcting code framework of [BH25], we use GPT-6 to find a nondeterministic counterexample based on a new choice of Reed--Solomon local codes, and then convert this example into a causal code using an edge orientation and local search.

0.39

Peeling Half the Onion: Embedding $k$-Outerplanar Graphs into $\ell_1$ and Trees with a Polynomial Distortion (in $k$)

cs.DS cs.CG ·2026-10-07 ·Hsien-Chih Chang, Jonathan Conroy, William Eliot, Hung Le, Vinayak
publishedtheory
Abstract · 22 pages, 2 figures, SODA'27

Chekuri, Gupta, Newman, Rabinovich, and Sinclair [SODA'03] showed that $k$-outerplanar graphs can be embedded into trees and $\ell_1$ with distortion $2^{O(k)}$. Their result is perhaps the strongest evidence supporting the still-open planar $\ell_1$-embedding conjecture: planar metrics can be embedded into $\ell_1$ with constant distortion. However, the exponential dependency on $k$ in their distortion bound remains the state of the art. Their embedding is obtained by a so-called onion-peeling approach: removing one layer of the graph at a time at the cost of incurring $O(1)$ distortion multiplicatively, and recursively embedding the resulting $(k-1)$-outerplanar graph. In this paper, we devise a new onion peeling approach that peels $k/2$ layers off the input $k$-outerplanar graph at every step. As a result, we obtain the first embedding of $k$-outerplanar graphs into trees and $\ell_1$ with distortion $\operatorname{poly}(k)$, an exponential improvement over the best-known distortion bound [SODA'03]. Our result of embeddings into trees comes closer to the distortion lower bound $Ω(k)$ for $k$-outerplanar graphs.

0.38

A deterministic polynomial-time algorithm for the Bilu--Linial signing problem on bipartite graphs

cs.DS cs.CCmath.CO ·2026-10-07 ·Zhiqiang Xu
Abstract · 20 pages. We have improved the presentation and clarified some details of the arguments

We give a deterministic polynomial-time algorithm for the Bilu--Linial signing problem on bipartite graphs. For every finite simple bipartite graph of maximum degree at most an integer $Δ\ge3$, the algorithm assigns signs to its edges so that the signed adjacency matrix has operator norm strictly less than $2\sqrt{Δ-1}$. Our algorithm builds on the randomized recursive repair framework of Jadbabaie, Saberi, and Sra~\cite{JSS26}, with deterministic rules for sign selection and vertex deletion.

0.37

q-ary GRAND

cs.IT ·2026-10-07 ·Reza Dianat, Majid Nasiri Khormuji, Hiroyuki Seki
published
Abstract · Accepted, IEEE Comm Letters

We develop q-ary Guessing Random Additive Noise Decoding (GRAND) for linear codes over a q-ary alphabet. The decoder works on a sorted symbol-likelihood array and uses three local child generation rules to generate symbol deviation patterns. We then prove that these rules induce a monotone rooted spanning tree of the full row-index space, so best-first traversal gives maximum-likelihood (ML) decoding under unlimited search. Reed-Solomon simulations verify ML agreement and show the finite-budget performance-complexity tradeoff.

0.36

Beyond LLM-GA: Secure Fluid Antenna Systems with ReEvo-Designed Memetic Algorithm

cs.IT cs.AI ·2026-10-07 ·Hanyong Xu, Zhaolai Dang, Tong Zhang
published
Abstract · Accepted by WCSP 2026

Fluid antenna systems (FASs) offer significant spatial flexibility, yet securing them against eavesdropping is critical for practical FAS deployment in military, satellite, and internet-of-things networks. Although large language model (LLM)-assisted genetic algorithms (LLM-GAs) can address this secure FAS port selection problem, whether further algorithmic improvement is possible warrants deeper investigation. To this end, we propose a memetic algorithm based on reflective evolution (ReEvo). Unlike the state-of-the-art LLM-GAs, which design only crossover or mutation operators with an LLM, our algorithm leverages an LLM to evolve dedicated crossover, mutation, and local-search operators offline. These operators are then embedded into a memetic search framework, thereby obviating any online LLM queries during execution. Simulation results at equal generation counts demonstrate that our proposed algorithm achieves a higher secure sum-rate than the conventional GA and the state-of-the-art LLM-GAs.

0.36

Subset Sum via Partial Match

cs.DS cs.CC ·2026-10-07 ·Baitian Li, Lixi Ye
Abstract · 33 pages, 1 figure

We show that the subset sum problem can be solved in randomized time $O(2^{0.49999\overline{7}n})$, breaking the $2^{n/2}$ meet-in-the-middle barrier. Our approach builds on the representation technique framework of Randolph and Węgrzycki (STOC 2026). We choose a representation for which testing the compatibility of pairs of partial solution vectors is exactly the partial match problem. To beat exponent $1/2$ for subset sum, it then suffices to give a nontrivial partial match algorithm in a certain parameter regime. We achieve this by designing a depth-2 linear circuit for the partial match matrix, which yields an efficient algorithm via the framework of Alman and Li (FOCS 2025).

0.34

Barely Monotone (min,+)-Convolution in Truly Subquadratic Time

cs.DS ·2026-10-07 ·MohammadTaghi Hajiaghayi, Danny Mittal, Saeed Seddighin
theory
Abstract

The (min,+)-convolution of two sequences A and B of length n is the sequence C with C[k] = min_{i+j=k} (A[i]+B[j]). For bounded inputs, whose entries are integers in {0,...,O(n)}, prior work computes it in truly subquadratic time when the inputs are monotone; the algorithm of Chi, Duan, Xie, and Zhang (STOC 2022) takes expected O~(n^{1.5}) time. We introduce a monotonicity measure ranging from 0 (monotone) to 1/2 (entirely non-monotone): a sequence has monotonicity alpha if it can be partitioned into O(n^alpha) monotone subsequences, and by the Erdos-Szekeres theorem every sequence has monotonicity at most 1/2. We show that truly subquadratic time is achievable even when just one input is barely monotone, that is, has monotonicity 1/2 - Omega(1): if A has monotonicity alpha, we compute the convolution in expected time O~(n^{5/3+2alpha/3}) for every bounded B. If B has monotonicity beta as well, the expected time improves to O~(n^{(3+alpha+beta)/2}), which matches the monotone case for alpha = beta = 0; this algorithm also allows infinite entries placed arbitrarily. We complement these algorithms with fine-grained reductions. Bounded (min,+)-convolution reduces to bounded monotone (min,+)-convolution of length N = O(n^{1.5}), so an O(N^{4/3-eps})-time algorithm for monotone inputs would give an O(n^{2-3eps/2})-time algorithm for bounded inputs. Similarly, entries bounded by n reduce to entries bounded by N^x on sequences of length N = Theta(n^{2/(1+x)}). We also show that if only A has entries in {0,...,M}, we can compute the convolution in O~(n(M+1)) time, and in O~(n^{1.5} sqrt(M)) time if A may also contain +infinity.

0.34

Non-Clairvoyant Scheduling is Hard Even for Trees

cs.DS ·2026-10-07 ·Kunal Agrawal, Owen Druzgal, Milind Prabhu, Jinhao Zhao
theory
Abstract

We study online scheduling of parallel jobs on $m$ identical processors to minimize maximum flow time. Each arriving job is represented by a directed acyclic graph (DAG) whose vertices are unit-time subjobs and whose edges specify precedence constraints. We consider non-clairvoyant algorithms: the DAG is not known when a job arrives, and each subjob is revealed only when it becomes ready. Agrawal, Moseley, Newman, and Pruhs (SPAA 2024) showed that First-In-First-Out (FIFO) has competitive ratio $Ω(\log m)$ even when every job is an out-tree. They also proved that FIFO is $O(\log m)$-competitive in several natural settings and asked whether this guarantee extends to general instances. More broadly, they asked whether any non-clairvoyant algorithm can be $O(1)$-competitive. We answer both questions in the negative by proving a lower bound of $Ω(\min\{m,\mathrm{OPT}\})$ for every non-clairvoyant online algorithm, where $\mathrm{OPT}$ is the maximum flow time of an optimal offline schedule. In particular, every non-clairvoyant algorithm has competitive ratio $Ω(m)$ on some instance with $\mathrm{OPT} \ge m$. The lower bound holds even when every job is an out-forest. We complement these lower bounds with an asymptotically optimal non-clairvoyant algorithm. FIFO is $O(m)$-competitive, but can have competitive ratio $Ω(m)$ even when $\mathrm{OPT}=O(1)$. For instances with small $\mathrm{OPT}$, we design a Small-Frontier-First algorithm that is $O(\mathrm{OPT})$-competitive. Combining the two algorithms yields an $O(\min\{m,\mathrm{OPT}\})$-competitive non-clairvoyant algorithm.

0.34

Physics-Aware Power Optimization for Rydberg Quantum Arrays via Barankin Bound Minimization

cs.IT ·2026-10-07 ·Yixiao Li, Jiayi Zhang, Bokai Xu, Yin Zhang
Abstract · 7 pages, 5 figures, IEEE GlobeCom 2026

Rydberg-atom quantum uniform linear arrays (RAQ-ULAs) offer a promising sensing architecture for direction-of-arrival (DOA) estimation in terahertz (THz) beam alignment. Existing studies often model the quantum receiver as a macroscopic linear block and rely on the Cramer-Rao bound (CRB) for array evaluation or power allocation. However, as a local variance bound, the CRB cannot capture threshold breakdown caused by spatial ambiguities in low-signal-to-noise-ratio (SNR) regimes. This paper proposes a physics-aware power optimization framework for RAQ-ULAs based on stochastic Barankin bound minimization. We derive a closed-form stochastic multipoint Barankin bound (BRB) matrix under a low-SNR integrability condition and introduce an arcsine-prior calibration to characterize bounded-domain error saturation. By incorporating a Lindblad-guided electromagnetically induced transparency (EIT) readout model, we couple the ambiguity-sensitive BRB with photon shot noise, power broadening, and laser Rabi frequencies. A CRB-optimized analytical baseline is further derived to expose the limitation of local-SNR-based allocation. To solve the resulting nonconvex problem, we develop a backtracking majorization-minimization (MM) algorithm with a Lipschitz-based quadratic surrogate. The algorithm ensures monotonic decrease and converges to a feasible stationary point. Simulations show that the proposed framework predicts threshold breakdown more accurately than the CRB and enlarges the reliable operating region under severe THz attenuation.

0.33

Symmetric Submodular Minimization from Comparisons

cs.DS ·2026-10-07 ·James Fox, David P. Woodruff
theory
Abstract

Given value-oracle access to a symmetric submodular function $f:2^V\to\mathbb{R}$ with $|V|=n$, a nontrivial minimizer can be found using $O(n^3)$ value queries. We study the weaker comparison model, in which a query on $S,T\subseteq V$ reveals only whether $f(S)$ is smaller than, equal to, or larger than $f(T)$. We give a deterministic polynomial-time algorithm that finds a nontrivial minimizer of any symmetric submodular function using $O(n^3)$ comparisons, matching the best-known deterministic value-oracle bound despite not knowing the function values. More generally, the same $O(n^3)$-comparison bound holds for minimization over the nonempty members of any downward-closed family. Our algorithm combines the minimum-capacity ordering recently introduced by Iwata and Konno with the contraction framework of Goemans and Soto. Applying this result to weighted graph cut functions resolves the main open question of Cohen-Addad et al., who gave an $\widetilde{O}(n^3)$-comparison algorithm that runs in exponential time and asked whether a weighted minimum cut can be found in polynomial time using comparisons. For graphs with $m$ edges of integer weight at most $B$, we also give a deterministic polynomial-time algorithm that finds a minimum cut using \[ \widetilde{O}\!\left(n^2+\min\!\left\{mB,\,nB^2\right\}\right) \] comparisons, improving on the $O(n^3)$ bound when $B$ is small. Finally, we show that every randomized algorithm that outputs a minimum cut with probability at least $2/3$ makes $Ω(n \log n)$ expected comparisons in the worst case. Under the stronger assumption that all edge weights are polynomially bounded integers, we obtain an $Ω(n \log \log n)$ expected comparison lower bound. These bounds contrast with the value-oracle model, where no $ω(n)$ lower bound is known even for deterministic algorithms.

0.33

Risk-minimizing GUE-implementations of truthful reporting in a Dynamic Social Choice Problem

cs.GT econ.THmath.COmath.PR ·2026-10-07 ·Endre Csóka, Máté Viczián
Abstract

The Guaranteed Utility Mechanism (GUM) was introduced to implement efficiency in Guaranteed Utility Equilibrium (GUE) in dynamic social choice problems. We show that this mechanism also has a previously unknown connection to risk minimization. Fixing an efficient decision policy, we study variance-minimizing budget-balanced transfer rules in the class of GUE-implementations of truthful reporting. A symmetrized version of GUM (sym-GUM) minimizes the sum of the agents' utility variances under truthful reporting when there is no public state. When a public state is present, a slight modification of this rule yields the minimizer for the same objective. Sym-GUM, however, does not minimize transfer risk. The minimizer for that problem is instead the symmetrized version of a closely related GUM-like transfer rule. In both problems, the minimizing truthful total transfers are unique up to agent-specific constants summing to zero and events of probability zero.

0.33

Almost Optimal Constant-Round Approximation of Dominating Set in Graph Classes with Excluded Minors

cs.DS cs.DCcs.DM ·2026-10-07 ·Saeed Amiri, Sebastian Siebertz
theory
Abstract

For every fixed proper minor-closed class $\mathscr C$ and every $ε>0$, we give a deterministic LOCAL algorithm that returns a dominating set of size at most $(2a(\mathscr C)+1+ε)γ_f(G)$ on every $G\in\mathscr C$. Here $a(\mathscr C)$ is the supremum edge-to-vertex ratio in $\mathscr C$, and $γ_f(G)$ is the fractional domination number. The class also admits a deterministic $(1+ε)$-approximation for fractional dominating set and a randomized algorithm that always returns a dominating set and has expected size at most $(1+ε)γ(G)$. In each case, the number of rounds depends only on $\mathscr C$ and $ε$. None of these algorithms requires the number of vertices or the maximum degree as part of the input. For planar graphs, this gives the deterministic guarantee $(7+ε)γ_f(G)$. Together with the lower bound of Hilke, Lenzen and Suomela, it determines the infimum of the deterministic constant-round approximation ratios for planar minimum dominating set as $7$, settling a question that had remained open since their work. The corresponding infima, measured against the integral optimum, are $7$ for graphs of Euler genus at most any fixed $g\ge0$, $2t-3$ for $K_t$-minor-free graphs with $3\le t\le9$, and $2r+1$ for graphs of treewidth or pathwidth at most any fixed $r\ge1$. We also prove that, for every integer $r\ge1$, no deterministic constant-round LOCAL algorithm achieves an approximation ratio below $2r+1$ on the $r$-th powers of paths, even when every vertex knows the number of vertices. This gives a new proof that the limiting constants are optimal for planar graphs, graphs of bounded treewidth or pathwidth, and $K_t$-minor-free graphs with $3\le t\le9$. For triangle-free planar graphs, the corresponding infimum is $5$.

0.32

Function-Correcting Lee-distance Codes for Symbol-Pair Read Channels

cs.IT ·2026-10-07 ·Hareesh K., B. Sundar Rajan
theory
Abstract · 27 pages; A short version to be communicated to ITW27

Function-correcting codes (FCCs) protect specified function evaluations of messages against errors while reducing the redundancy required for reliable communication. We study FCCs for symbol-pair read channels, motivated by high-density storage systems that read overlapping symbol pairs. Phase-shift keying (PSK) modulation is well-suited to such systems due to its bandwidth efficiency and noise robustness. While FCCs for symbol-pair read channels have been studied under the Hamming metric, the Lee metric is a more appropriate error model for $q$-ary PSK and, hence for $q$-ary symbol-pair read channels. We generalize the symbol-pair Lee distance, previously defined only over $\mathbb{Z}_4$, to $\mathbb{Z}_q$, $q\ge2$, and introduce function-correcting symbol-pair Lee-distance codes (FCSPLCs) over $\mathbb{Z}_q$, specializing to $q=2^m$, $m\ge1$, for $2^m$-ary PSK constellations. We investigate their redundancy requirements by introducing irregular-pair Lee-distance codes and relating the optimal redundancy of FCSPLCs to the shortest length of such codes. We derive Plotkin-type and Gilbert--Varshamov-type bounds on the optimal redundancy through lower and upper bounds on the shortest length of irregular-pair Lee-distance codes. For bijective functions, we obtain corresponding Plotkin-type and Gilbert--Varshamov-type bounds for classical Lee metric codes for symbol-pair read channels over $\mathbb{Z}_q$, which, to the best of our knowledge, are the first such bounds for the Lee metric symbol-pair setting. We then specialize the FCSPLC framework to pair-locally bounded functions, the Pair-Lee weight function, and the Pair-Lee weight distribution function, giving explicit constructions and corresponding bounds on the optimal redundancy. Finally, for linear functions, we derive a Plotkin-type lower bound on the optimal redundancy.

0.32

Primal/Dual Method for the Grothendieck Constant

cs.CC math.FA ·2026-10-07 ·Steven Heilman, Chris Jones, Giulio Malavolta
Abstract

We determine the hundredths digit of the real Grothendieck constant by proving $1.773 \leq K_G \leq 1.7799$. Furthermore, numerical heuristics suggest the estimate $K_G \approx 1.779$. We start by simplifying the Gaussian duality theory of $K_G$, based on Hermite projection games and Krivine rounding schemes. The bounds are obtained by a heuristic primal/dual search for these objects, followed by rigorous {\em certification} of their values. The search and certification are performed using AI tools, and in particular the rigorous certificates are very large (analytical reductions to thousands of numerical inequalities).

0.32

Tightness and Error Exponents of SDP with Logarithmically Many Communities

cs.IT math.ST ·2026-10-07 ·Shengtao Guo, Ethan X. Fang, Junwei Lu
theory
Abstract · 36 pages, 8 figures

We study a semidefinite programming (SDP) relaxation for community recovery when the number of communities grows logarithmically. In the balanced stochastic block model with $n=km$ vertices, we consider the regime $k/\log m\toγ>0$, with edge probabilities $α\log m/m$ within communities and $β\log m/m$ across them, for fixed $α>β>0$. We derive the sharp asymptotic tightness boundary away from critical cases. When rare vertices cause tightness to fail while the bulk remains spectrally stable, the normalized matrix error of every near-optimal solution still vanishes. We prove matching high-probability exponents for this error and the normalized optimal objective gain, governed by the same local correction that determines tightness. Throughout this spectrally stable region, a single SDP solve followed by explicit rounding and refinement achieves exact community recovery above the information-theoretic threshold. This guarantee holds even when the planted community matrix is not an optimal solution to the SDP.

0.31

Generating cyclic pivot Gray codes for well-ordered $k$-degenerate graphs in constant amortized time

cs.DS ·2026-10-07 ·Lei Dong, Dennis Wong, Bowie Liu, Rui Bao, Lin Chen +2 more
Abstract · 38 pages, 4 figures

A graph $G$ is $k$-degenerate if there exists an ordering $v_1, v_2, \dots, v_n$ of its vertices such that each vertex $v_i$ has at most $k$ neighbors $v_j$ in $G$ with $j < i$. A well-ordered $k$-degenerate graph is a labeled graph on the vertex set $\{1, 2, \dots, n\}$ in which every vertex $i$ has at most $k$ neighbors among $1, 2, \dots, i-1$. We present the first simple algorithms that generate, rank, and unrank cyclic pivot Gray codes for well-ordered $k$-degenerate graphs, where consecutive graphs differ by the addition, removal, or pivoting of a single edge. Our algorithm generates each well-ordered $k$-degenerate graph in constant amortized time per graph, using $O(n^2)$ space, while ranking and unranking take $O(n^2)$ time and $O(n^2)$ space.

0.30

Sharper bounds on the thresholds for sensitive and Blackwell optimality

cs.GT ·2026-10-07 ·Xavier Allamigeon, Stéphane Gaubert, Julien Grand-Clément, Ricardo Katz
Abstract

In perfect-information two-player stochastic games, the notions of Blackwell and sensitive optimality provide generalizations of the classical mean-payoff optimality and discount optimality criteria to account for more farsighted preferences. We provide bounds on the Blackwell threshold $α_{\sf bw}$ and the $d$-sensitive thresholds $α_{\sf d}$, defined as the smallest discount factors above which discount optimal policies coincide with Blackwell optimal policies and $d$-sensitive optimal policies respectively. Our refined bounds improve upon prior work by focusing on ``reduced'' families of polynomials and, crucially, our bounds are tight in terms of controlling the degrees and heights of the minimal polynomials of the Blackwell thresholds. We apply classical root separation based on Mahler's and Cauchy's bounds to our reduced families to derive the strongest upper and lower bounds on $α_{\sf bw}$ in the literature, and we are the first to obtain bounds on $α_{\sf d}$ in the multichain stochastic setting.

0.30

The Confidence Game: Strategic Miscalibration in Human-AI Delegation

cs.GT cs.AIcs.CLcs.CY+1 ·2026-10-07 ·Raghu Arghal, Saswati Sarkar, Shirin Saeedi Bidokhti
Abstract

Calibrated uncertainty quantification is essential to ensuring AI agents are trustworthy and reliable. However, when agents seek to maximize user engagement or revenue, confidence reports may be strategically distorted, detracting from their informativeness. We formalize this problem in the Confidence Game: a repeated signaling game with imperfect monitoring in which an agent of unknown honesty and ability reports its confidence, and a user decides whether to delegate the task or complete it herself. The agent manages the tradeoff between manipulating signals and maintaining its reputation. We characterize the Markov Perfect Bayesian Equilibria of the two-period game and show that honest reporting is not an equilibrium, inflation is the unique best response once the agent is sufficiently myopic, and under-reporting requires that the user believe honesty to be a minority. We then place an LLM in the agent role, supplying it with its true probability of success so that any gap between what it knows and what it reports is attributable to incentives rather than to miscalibration. The model claims high confidence on 56% of tasks it has been told it will probably fail. This persists on real tasks, where it must estimate its own accuracy and causes miscalibration to increase while the agent's signal becomes less informative. Furthermore, we find that the LLM agent's decisions are coherent, but it systematically underestimates both how likely the user is to delegate and how secure its reputation is, resulting in less extreme behavior. Pricing the agent's reporting rule, we find that it destroys 68% of the gains from delegation, of which 71% is information the report no longer carries and no amount of user sophistication recovers. Overall, we establish confidence reporting under delegation as a strategic problem and provide a tractable basis for modeling, analyzing, and testing agent behavior.

0.29

The Economic Security of Exponential EIP-1559

cs.GT cs.CRcs.DC ·2026-10-07 ·Ben Berger, Edward W. Felten, Robin Fritsch
theory
Abstract · 29 pages, 2 figures; includes appendices

We consider the problem of parameter selection for EIP-1559, a widely used gas pricing mechanism for blockchains. As opposed to previous work, we aim to achieve economic security, where the chosen parameters guarantee a lower bound on total collected gas fees whenever usage over a given time window exceeds a specified threshold. This bound can be set prohibitively high, thereby economically deterring usage above a desired level. Such guarantees are particularly relevant to long-term objectives such as limiting state growth. We apply our approach to the pure exponential version of EIP-1559 which Ethereum uses and to a variant used by Robinhood Chain and Arbitrum. To achieve economic security, we first characterize the revenue-minimizing gas-usage distribution for the exponential version and for the other variant we construct a distribution whose associated fee revenue provably approximates the minimum. Using these results, we then demonstrate how to set economically secure mechanism parameters for both variants and we discuss the associated tradeoffs.

0.29

The smallest programmable machine and the hardness of analyzing it

cs.CC ·2026-10-07 ·J. Andres Montoya
theory
Abstract

We investigate the computational complexity of analyzing the structural and behavioral properties of deterministic k-pebble automata, which represent a natural framework for studying minimal programmable machines. First, we provide an explicit construction of a three-pebble automaton U capable of simulating any deterministic finite automaton (DFA) on a given input word, establishing a link between pebble automata capabilities, Kolmogorov complexity, and automatic complexity. Next, we restrict our architectural framework to two-pebble programmable machines U(2,N) simulating DFAs with at most N states. We analyze the decision problem ANAL(U(2,5)), which asks whether a five-state program can separate two distinct finite words. By establishing a polynomial-time reduction from the quasi-identity checking problem for finite semigroups, we prove that ANAL(U(2,5)) is NP-complete. Finally, we explore the complexity of separating binary strings under the Exponential Time Hypothesis (ETH), showing how computational hardness implies polynomial lower bounds for the size of the smallest DFAs separating binary strings.

0.29

Decomposition Profiles and Weisfeiler-Leman Dimension for Graphs of Bounded Rank Width

cs.DM math.COcs.LO ·2026-10-07 ·Antonios Kalampakas
theory
Abstract · 44 pages. Includes ancillary verification code

The Weisfeiler-Leman (WL) algorithm tests graph isomorphism by assigning colours to $d$-tuples of vertices and iteratively refining the colouring according to neighbourhood configurations until it stabilizes. Increasing $d$ allows the test to detect finer structural differences between graphs, but the computational cost grows exponentially with $d$. The central question is which dimension allows us to distinguish nonisomorphic graphs. We relate a sufficient such dimension to the structure of a graph $G$ through a rank decomposition over $\mathbb F_2$. The two parameters governing the bound are the maximum cut rank $w$ and the fork load $\ell$, which is the largest sum of the parent and two child cut ranks at a fork in the decomposition tree. We prove that WL in dimension $\max\{\ell+1,2w+2\}$ distinguishes every finite nonempty vertex-coloured graph $G$ with such a decomposition from every graph not isomorphic to $G$. Consequently, for $k\geq1$, graphs of rank width at most $k$ can be distinguished from every nonisomorphic graph in dimension $3k+1$, graphs of linear rank width at most $k$ in dimension $2k+2$, and edgeless graphs in dimension one. Uncoloured Cai-Fürer-Immerman (CFI) constructions give linear lower bounds for both width parameters, so the upper bounds are tight up to constant factors.

0.29

Asymptotically Optimal Public Project Redistribution without Bounded Precision: A Complex-Analytic Proof

cs.GT ·2026-10-07 ·Mingyu Guo
Abstract

We study worst-case VCG redistribution mechanism design for the public project problem, where n agents decide whether to build a non-excludable public project and the designer chooses a Groves term that returns as much of the VCG payment as possible without running a deficit. The objective is the worst-case efficiency ratio, the worst-case ratio between the agents' total utility and the first-best total utility. Prior work showed that this ratio can approach 1 as n grows only under a bounded precision assumption, that all types are rational numbers with a common bounded denominator. The assumption is stronger than it looks: it confines the difficult profiles to those in which all but a bounded number of agents report zero, and an agent who reports zero knows the total exactly. Without it, the best known guarantee tends to 1/2. We remove the assumption. We construct a deterministic, anonymous, strategy-proof, and non-deficit mechanism for arbitrary real types in [0,1] whose worst-case efficiency ratio is 1-O(1/log log n). The construction estimates the shortfall of the reported total below the project cost by filling in each agent's missing report with a resampled one, then estimates the error of that estimate, then the error of the error, and so on. After m rounds the aggregate error is at most 1/Hm mean reports, where Hm is the mth harmonic number. The proof of this bound is the heart of the paper and, unusually for a result about payments, uses complex analysis: a trigonometric representation of the shortfall function together with the Poisson-Jensen inequality for bounded analytic functions in the upper half plane.

0.29

A $G$-Invariant Erdős--Hanani Theorem with Applications to Optical Orthogonal and Constant-Weight Codes

cs.IT math.CO ·2026-10-07 ·Yeow Meng Chee
theory
Abstract · 16 pages, 1 table

Let $k>t\ge 2$ be fixed integers, and let $G$ be a permutation group on a set of $v$ points. A $G$-invariant packing is free if every block has $|G|$ distinct images under $G$. We give two conditions on $G$ under which, as $v\to\infty$, there is a free $G$-invariant $t$-$(v,k,1)$ packing with $(1-o(1))\binom{v}{t}/\binom{k}{t}$ blocks. This is a $G$-invariant form of the Erdős--Hanani theorem on asymptotically optimal packings. For semiregular groups $G$, packings with these properties exist if and only if all but $o(v^t)$ of the $t$-subsets of points have trivial setwise stabiliser. This property is satisfied by every semiregular group for $t\ge 3$, and by every cyclic semiregular group for $t=2$. As a consequence, the Johnson bound is asymptotically attained at all lengths by optical orthogonal codes of constant weight $w$ and constant correlation $λ$ with $w\geλ+2$, by their multidimensional and signature pattern versions, and by cyclic and quasi-cyclic constant-weight codes. The same method gives a form of the theorem for group divisible packings that are invariant under a cyclic group. This determines the asymptotic maximum size of two-dimensional optical orthogonal codes with at most one pulse per wavelength.

0.28

$\exists \mathbb{R} \subseteq \textsf{CH}$

cs.CC cs.CG ·2026-10-07 ·Alex Meiburg
Abstract · 114 pages

The existential theory of the reals asks whether polynomial constraints with integer coefficients have a real solution. We give a proof placing this problem in the counting hierarchy. The first argument is intended to expose the essential steps, and a separate analysis lowers the bound to $\exists \mathbb{R}\subseteq\textsf{BPP}^{\textsf C_3\textsf P}\subseteq\textsf C_4\textsf P$, the fourth level of the hierarchy. For each fixed $w$, sentences with $w$ alternating real quantifier blocks lie in $\textsf C_{9w+17}\textsf P$. We then treat exact semidefinite feasibility, PosSLP, square-root sum, geometric real counting, Euler characteristic, and complex feasibility in separate applications. The corresponding bounds include $\textsf{BPP}^{\textsf C_2\textsf P}$ for general SDP, $\textsf{BPP}^{\textsf{PP}}\cap\textsf{P}^{\textsf{NP}^{\textsf{PP}}}$ for PosSLP and square-root sum, and $\textsf{FP}^{\textsf C_4\textsf P}$ for total geometric real counting. Note: These proofs were discovered by ChatGPT after a series of conversations ending on September 29th 2026. A group of researchers has been working to digest the proof, and while the most essential arguments appear correct, we are endeavoring to give this result the treatment it deserves and a proper exposition and development to benefit of the community. However, on October 6th, OpenAI released a very similar result, with a slightly weaker bound. While we work to improve our exposition of this proof, the current version has been uploaded as a service to the community to compare the different proof techniques. While the listed author takes responsibility that the proofs appear to be correct, he has not played a nontrivial role in developing them, and believes the human value will be in good exposition and canonicalization of the results.

0.28

Optimal Regret for Online Market Making with Limit Order Book

cs.GT cs.LG ·2026-10-07 ·Maria Elena Vischi, Francesco Emanuele Stradi, Alberto Marchesi
hot direction
Abstract

We study online learning in market making, where, at each round, a market maker posts bid and ask prices before observing the market price and the private valuation of an incoming trader. In this setting, Maran et al. 2026 introduce a feedback model motivated by limit order books, in which the trader's valuation is revealed only if no transaction occurs. Assuming that trader valuations are drawn i.i.d. from an unknown distribution while market prices are chosen adversarially, they establish an expected regret bound of $\widetilde{\mathcal{O}}(T^{2/3})$. In this work, we improve upon this guarantee by establishing a high-probability regret bound of $\widetilde{\mathcal{O}}(\sqrt{T})$. As a warm-up, we first consider the full-feedback setting. We introduce a discretization of the bid-ask space based on two coupled grids and combine it with Hedge to achieve the desired regret rate. Building on these ideas, we then address the substantially weaker feedback induced by a limit order book and develop an algorithm that achieves the same guarantee. Finally, we investigate the limits of learnability in fully adversarial environments, where the valuations may vary arbitrarily as well. Perhaps surprisingly, we show that when both market prices and trader valuations are chosen adversarially, sublinear regret is impossible even under full feedback, thereby motivating our stochastic assumption on the valuations.

0.28

Beyond Nominal Equilibria: Risk-Averse Multi-Population Mean-Field Games

cs.GT math.OCcs.LGeess.SY ·2026-10-07 ·Bhavini Jeloka, Siddhartha Ganguly, Panagiotis Tsiotras
Abstract · Submitted to a conference; comments are welcome

Recent advances in mean-field games and its multi-population variants enable large-scale heterogeneous multi-agent systems to be modeled through representative agents and their associated mean-field distributions. However, existing approaches do not explicitly account for uncertainty in the behavior of other populations. To this end, we introduce a new paradigm: risk-averse multi-population mean-field games, where each population optimizes a worst-case expected reward over dynamically feasible ambiguity sets of mean-field flows of a subset of the other populations. Employing an occupation-measure formulation along with tools from set-valued analysis, we establish, under mild assumptions, several theoretical properties of the multi-population game, including the geometric properties of the ambiguity sets and the existence of a novel risk-averse multi-population mean-field equilibrium. Further, we derive contractivity results of the fixed-point operator under entropy regularization and show that it can be utilized to learn the equilibrium. Finally, we propose a risk-averse fictitious-play scheme and show that exploitability decays to zero, despite the additional nonlinearity introduced by the worst-case objective. We report several numerical experiments to illustrate convergence and risk-averse behavior.

0.28

Asymptotic Subsidies for Envy-Free Allocation

cs.GT ·2026-10-07 ·Zekai Wu
published
Abstract · Accepted by IJTCS-FAW 2026

We study the minimum monetary subsidy required for exact envy-freeness in allocating $m_n$ indivisible goods to $n$ agents. Valuations are additive, item values are i.i.d. from a distribution on $[0,1]$ with density bounded above and away from zero, and all agents value money equally. We consider $n\to\infty$ with $m_n=qn+r_n$, where $q\ge0$ is a fixed integer and $0\le r_n<n$. For the nonzero-remainder case where $1\le r_n<n$, we give a cubic-time algorithm based on a maximum-weight matching of balanced bundles and complementary dual prices. It returns an envy-free outcome for every instance. Under the random model, both its payment and the unrestricted minimum are $n-r_n+o_p(n)$ for $q\ge1$. When $q=0$ and $r_n/n$ converges to a limit below one, the error term improves to $O_p(1)$. For exact divisibility, where $r_n=0$, prior work gives zero subsidy with high probability for $q\ge2$. The square case $q=1$ is exceptional. Its minimum subsidy is $Θ_p(\log n)$ and, with high probability, equals the optimum over allocations in which every agent receives exactly one good. This restricted optimum can be computed in cubic time using a maximum-weight perfect matching and all-pairs shortest paths. Together, these results provide a unified asymptotic characterization of the minimum subsidy.

0.27

Building codes with transversal CCZ using projective geometry and SAT solvers

cs.IT quant-ph ·2026-10-07 ·Bohan Lu, Kenneth R. Brown
theorycode/data
Abstract · 40 pages, 3 figures. Code and data: https://github.com/jerrylvx/gate2code-public

We construct CSS codes with three logical qubits whose logical CCZ gate is implemented by transversal physical $T/T^\dagger$ gates and whose distance is at least 3. These codes start with 8-divisible $X$-stabilizers defined by projective geometry, which guarantees $d_Z \geq 3$. A SAT solver then finds three $X$-logical operators compatible with the transversal CCZ. We build codes at thirteen block lengths from $n=48$ to $n=496$, with $d_Z=3$ or 4, and illustrate the method with the [[48,3,3]] code $Q_{48}$. Additionally, we prove that no CSS code with three logical qubits, $Z$-distance at least 3, and a quasi-transversal CCZ gate (transversal $T$ with a diagonal Clifford correction) exists below $n=39$. The [[47,3,3]] code of Jacinto et al. attains the minimum length in this class, as established by the classification of Wills, Jain, and Singh.

0.27

Breaking the $\sqrt{3}$ Barrier for Maximum Weighted $3$-Set Packing

cs.DS ·2026-10-07 ·Weitian Tong, Yao Xu
theory
Abstract · 38 pages, 6 figures

We give a deterministic polynomial-time $1.6908$-approximation for Maximum Weighted $3$-Set Packing, breaking the $\sqrt3$ locality-gap barrier of squared-weight local search. The approximation ratio for this problem progressed from Berman's $2$ [Ber00] to Neuwohner's $2-\frac{1}{63{,}700{,}992}+ε$ [Neu21]. Thiery and Ward then obtained $1.786$ [TW23], while Thiery subsequently improved the bound to $1.761+ε$ and finally to $\sqrt3 \approx 1.732051$ through a layered exchange analysis [Thi23]. Thiery also proved that $\sqrt3$ is a locality-gap lower bound for the squared-weight objective even with exchanges of arbitrary size. Our algorithm performs in two phases and combines two objectives. Phase~I computes a bounded-exchange local optimum for the squared-weight potential and analyzes it through Thiery's layered framework, while strengthening the terminal analysis by preserving internal tree-edge slack for nonsingleton components and exploiting the incidence structure of $3$-sets for final singletons. This yields an augmented structural inequality with residual positive claw gain under the original objective. Phase~II switches to the original objective and recovers sufficient residual gain through an auxiliary weighted $9$-Set Packing instance. A covering argument transfers the structural bound through the high-girth lift used only in the analysis.

0.27

Truly Sub-$3^n$ Min-Sum Subset Convolution and Join Ordering

cs.DS cs.DB ·2026-10-07 ·Mihail Stoian
hot direction
Abstract · Feedback welcome; v2: fixed technique naming

We present a deterministic reduction from min-sum subset convolution to min-plus matrix product. We show that if the min-plus product of two $D\times D$ matrices with $β$-bit integer entries can be computed in $D^{3-δ}\operatorname{poly}(β,\log D)$ time for a fixed rational $0<δ<1$, then min-sum subset convolution on an $n$-element universe can be solved in $(2+2^{-δ})^n 2^{O(\sqrt n\log(n+1))}\operatorname{poly}(n,β)$ time. Instantiating this reduction with the recent breakthrough on subcubic min-plus matrix product by Alman and Vassilevska Williams gives a Las Vegas algorithm with expected running time $O^*(2.9987^n)$ and a deterministic algorithm with running time $O^*(2.9997^n)$, strictly breaking the longstanding $3^n$ computational barrier. Notably, these speedups translate directly to database query optimization, yielding the same expected and deterministic running-time bounds for join ordering under the $C_{\mathrm{out}}$ cost function.

0.27

On the Cyclic Assumption of the Cow-Path Search Algorithm

cs.DS cs.LG ·2026-10-07 ·Yuan Ma, Yiqun Lisa Yin
Abstract

In the cow-path problem, a cow must find a goal lying at an unknown distance on one of $w$ paths connected only at the origin, and performance is measured by competitive ratio. Kao, Reif and Tate designed an efficient randomized algorithm in which the cow visits the paths in a fixed cyclic order. They proved the algorithm is optimal for $w=2$, and subsequently Kao, Ma, Sipser and Yin proved its optimality for all $w$, with a claim that no algorithm does better than the best cyclic one. This note provides a detailed proof of that claim.

0.27

0/1 integer programming over polytopes of Chvátal rank one is as hard as lattice problems

cs.CC math.OCcs.DM ·2026-10-07 ·Alberto Del Pia
Abstract

A rational polyhedron has Chvátal rank at most one if a single round of Chvátal-Gomory cuts yields its integer hull. In 1973, Chvátal asked whether 0/1 integer programs over polytopes of bounded Chvátal rank can be solved in polynomial time. For polyhedra of bounded Chvátal rank, integer feasibility is in NP $\cap$ coNP, so it is unlikely to be NP-hard. We answer Chvátal's question negatively already in its most restricted form: for Chvátal rank one, for integer feasibility, and even for the SAT polytopes of formulas whose clauses have at most three literals. Our answer holds under either of two standard assumptions from lattice-based cryptography. First, a polynomial-time algorithm for integer feasibility on such SAT polytopes of Chvátal rank one would solve bounded distance decoding with polynomial factors in deterministic polynomial time, contradicting the conjectured hardness of approximating lattice problems within polynomial factors. Second, assuming the hardness of learning with errors, an average-case analog of bounded distance decoding, the problem is also hard on average, for a polynomial-time samplable distribution of formulas. This also gives negative answers to related questions of Cornuéjols, Lee, and Li. We first prove these results for parallelepipeds, where they rest on an elementary sufficient condition: a polyhedron has Chvátal rank at most one if its width is less than one along every row of some unimodular matrix. For our parallelepipeds, such a matrix exists but is hard to find. For SAT polytopes, we represent candidate lattice vectors by their residues modulo small integers, and binary arithmetic on additional 0/1 variables checks that these vectors lie in the lattice.

0.27

Settling PROPm and PROPavg in Graphical Resource Allocation

cs.GT ·2026-10-07 ·Bo Li, Ankang Sun, Ruijie Wang
publishedtheory
Abstract · Full version of a paper to appear in WINE 2026. V2 includes an extended related work section

We study proportional fairness in graphical resource allocation, where agents are vertices, indivisible items are edges, and each item must be allocated to one of its two endpoints. It has been proved that PROP1 orientations always exist and PROPx orientations may not, but it has remained open whether the intermediate relaxations PROPm and PROPavg (both can be satisfied without graphical constraints) can always be satisfied. In this paper, we resolve this gap. We prove that PROPm orientations always exist for goods on multigraphs and can be computed efficiently. The guarantee extends to chores and a mixture of goods and chores. In sharp contrast, we show that PROPavg orientations need not exist, even for simple graphs with binary valuations, and that deciding their existence is NP-complete. We further quantify the efficiency loss of PROPm: for goods, the price of PROPm is exactly $2$, which is one of the few settings where a constant bound on the price of fairness can be obtained. We complement these results with hardness results for welfare optimization under PROPm and for the existence of orientations that are simultaneously PROPm and Pareto optimal.

0.25

Model-Order-Adaptive Channel Estimation for AFDM Systems with Fractional Delay and Doppler

cs.IT ·2026-10-07 ·Ruizhe Wang, Cunhua Pan, Hong Ren, Shaohua Li, Jiangzhou Wang
hot direction
Abstract

Affine frequency division multiplexing (AFDM) has emerged as a promising waveform for high-mobility communications owing to its ability to exploit multipath diversity. However, channel state information (CSI) acquisition remains challenging in channels with fractional delay and Doppler and an unknown number of propagation paths. In this paper, we investigate the channel estimation for AFDM systems. We first analyze the AFDM response in the presence of fractional-delay-induced frequency wrapping. The analysis reveals that fractional delay may displace the dominant extremum and generate informative secondary extrema. Then, we develop a model-order-adaptive channel estimator within an improved space-alternating generalized expectation-maximization (SAGE) framework. A dual-pilot reference symbol is employed for path and delay initialization, while pilot observations across multiple AFDM symbols provide temporal information for Doppler estimation. New paths are identified through statistically controlled residual tests, whereas unsupported or redundant paths are removed by conditional support pruning. The delay, Doppler frequency, and complex gain of each retained path are subsequently estimated through a coarse-to-fine procedure and refined using the exact AFDM likelihood. Simulation results demonstrate that the proposed method achieves lower normalized mean squared error (NMSE) and bit error rate (BER) than representative SAGE, sparse-recovery, and Bayesian benchmarks. It also provides reliable path detection and model-order estimation and maintains robust performance at terminal velocities of up to 600~km/h.

0.25

Marrying Pricing and Advertising with LLMs

cs.GT cs.LG ·2026-10-07 ·Alessandro Barro, Francesco Bacchiocchi, Francesco Emanuele Stradi, Alberto Marchesi
hot direction
Abstract

We study a sequential pricing problem in which a seller jointly posts a price and an advertisement generated by a large language model (LLM). The seller aims to maximize revenue under an unknown product demand that depends on both decisions, while observing only whether each offer leads to a purchase. We propose an online actor-critic algorithm that combines low-rank adaptation (LoRA) of a pretrained LLM with a demand model fitted to available data. At each round, the actor generates an advertisement, and the critic estimates purchase probabilities to guide price selection. Then, the resulting feedback is used to update both the actor and the critic, with the critic's revenue estimates providing a baseline for policy gradient updates of the actor. To evaluate our approach, we develop an evaluation framework with three synthetic demand models and a demand simulator built from real-world marketplace data. Finally, we compare our algorithm with benchmarks that do not jointly optimize price selection and advertisement generation, achieving expected revenue gains over the reference policy of 5.69%, 5.18% and 55.96% under the three synthetic demand models and 5.81% under the marketplace simulator.

0.25

Probabilistic GC-Constraints for Composite DNA

cs.IT ·2026-10-07 ·Humeyra Bodur, Frederik Walter, Antonia Wachter-Zeh
Abstract

This paper addresses the challenge of encoding biochemical GC-content constraints in composite DNA-based data storage. Previous deterministic models impose high rate penalties by avoiding any possibility for a strand to fall outside the allowed range. To account for the stochastic nature of composite DNA, we introduce an $ε$-probabilistic constraint framework, and derive capacity bounds for global constraints using composition types and for local sliding-window constraints via finite-state Markov chains. Furthermore, we propose a capacity-achieving multi-type enumerative encoder.

0.25

Polarforming for MIMO Covert Communications

cs.IT ·2026-10-07 ·Xin Xie, Jinpeng Xu, Yihang Huang, Li Zhou, Zhaolong Ning
benchmark
Abstract · 13 pages, 8 figures, Submitted for possible publication in an IEEE journal

Covert communication conceals wireless transmission activity, but conventional multi-antenna designs rely mainly on spatial beamforming and can be constrained by limited spatial degrees of freedom. This paper investigates polarization-reconfigurable antenna (PRA)-aided multiple-input multiple-output (MIMO) covert communication, where Alice and Bob perform transmit and receive polarforming, respectively, and Willie uses a vertically polarized antenna for radiometric detection. We jointly optimize the transmit covariance matrix and the transmit and receive phase shift vectors to maximize Bob's achievable rate under the covertness and transmit power constraints. Under Gaussian signaling, perfect polarized channel state information, and known noise power at Willie, we derive his optimal radiometric threshold and exact minimum detection error probability (DEP) for a finite observation length, and convert the DEP requirement into a deterministic constraint on his received signal power. We further prove that, for a fixed transmission design, positive signal leakage becomes detectable as the observation length tends to infinity. To solve the design problem for a finite observation length, we develop an alternating optimization (AO) algorithm with a generalized water-filling covariance update and closed-form phase updates. Numerical results validate the exact DEP analysis and the rapid convergence of the proposed algorithm. Compared with the sufficient divergence-based condition, the exact covertness constraint permits a $0.99$--$1.12$~dB higher received signal-to-noise ratio (SNR) at Willie, while joint transmit and receive polarforming provides a rate gain of up to $53.8\%$ over the considered benchmark schemes.

0.25

Packing Diverse Shortest Cycles

cs.DS ·2026-10-07 ·Akanksha Agrawal, Fedor V. Fomin, Petr A. Golovach, Vinod Gupta, Yash Hiren More +1 more
theory
Abstract

Bentert, Fomin, Golovach, Korhonen, Lochet, Panolan, Ramanujan, Saurabh, and Simonov (SODA 2025) initiated the parameterized study of Edge-Disjoint Shortest Cycle Packing: given a weighted graph $G$ and an integer $k$, decide whether $G$ contains $k$ edge-disjoint cycles of minimum weight. They showed that the problem admits an algorithm running in time $n^{O(k^6)}$ and asked whether it is fixed-parameter tractable or $W[1]$-hard parameterized by $k$. We resolve this question by proving that Edge-Disjoint Shortest Cycle Packing is $W[1]$-hard parameterized by $k$, even on unweighted subcubic graphs. The same lower bound also applies to the vertex-disjoint variant. For planar graphs, they provides a construction of a kernel with $O(k^2)$ vertices and an algorithm running in time $k^{O(k)} \cdot n^{O(1)}$, and explicitly asked whether the problem admits a single-exponential algorithm of running time $2^{O(k)} \cdot n^{O(1)}$. Rather than addressing this question in isolation, we introduce a more general framework, Diverse Shortest Cycle Coverage, which asks for $k$ shortest cycles that may overlap in a controlled way while maximizing the total weight of covered edges. This framework simultaneously captures edge-disjoint shortest cycle packing, the problem of finding diverse shortest cycles, and the problem of maximizing edge coverage by shortest cycles. Our main algorithmic result shows that Diverse Shortest Cycle Coverage can be solved on planar graphs in time $2^{O(k)} \cdot n^{O(1)}$, thereby giving a single-exponential algorithm for Edge-Disjoint Shortest Cycle Packing as a special case. The key idea of our algorithm is a structural analysis of the Laminar Shortest Cycles Tree, a tree-like decomposition that reveals a laminar interaction pattern among shortest cycles in planar graphs and enables an efficient dynamic programming algorithm.

0.24

Unconditional Boolean Dynamic Lower Bounds from Disjointness

cs.DS cs.CC ·2026-10-07 ·Young Kun Ko
theory
Abstract · 48 pages, 3 figures. v2: main result stated for the multiphase problem; appendix extended to further dynamic problems; introduction revised.…

We prove an unconditional $Ω((\log n/\log\log n)^2)$ lower bound on the maximum of the worst-case update time and expected query time for the insert-only version of Patrascu's multiphase Disjointness problem [STOC 2010]. The bound holds in the zero-error cell-probe model with logarithmic word size. Patrascu's reductions transfer it to directed reachability under edge insertions, already on directed acyclic graphs of depth $2$. This improves the $\widetildeΩ(\log^{3/2} n)$ bound of Larsen and Yu [SICOMP 2025] to the quadratic-logarithmic scale previously attained for dynamic range counting and multiphase Inner Product. The bound also transfers to a broad collection of dynamic graph, algebraic, geometric, and string problems. We also obtain a separation between amortized and worst-case time for streaming regular-expression membership. For a simple family of expressions of maximum length $M$ over a polynomial-size alphabet, a deterministic algorithm uses $O(1)$ amortized time per symbol on every input stream, whereas every zero-error cell-probe algorithm with word size $w=Θ(\log M)$ requires $Ω((\log M/\log\log M)^2)$ time on some symbol. The proof adapts Ko's 2.5-round multiphase communication framework [FOCS 2026] from Inner Product to Disjointness. Our main technical observation is that the protocol's verification step allows a one-sided corruption bound to replace the discrepancy property used for Inner Product.

0.24

Pathwise Information Certificates for Decentralized Adaptive Sensing

cs.IT cs.LG ·2026-10-07 ·Theodoros Tsiligkaridis
Abstract · 26 pages, preprint

We study decentralized adaptive sensing, where multiple agents choose measurements from evolving local beliefs while exchanging information over a communication graph. We ask whether the measurements actually selected by an adaptive policy have collected enough evidence to distinguish the true target from every plausible alternative. We develop a pathwise certificate based on the Rényi--Chernoff information accumulated along the realized sensing trajectory. It yields nonasymptotic MAP-error bounds and an anytime, network-wide stopping rule for arbitrary history-dependent sensing policies, while separating accumulated statistical information from a bounded network-mixing transient. Linear growth of the information against the least-resolved competitor implies exponential decay of MAP and squared-localization error. A classical pairwise KL converse, specialized to the adaptive decentralized transcript, shows that insufficient information on any pair prevents a positive uniform error exponent, confirming the hardest competitor as a fundamental bottleneck. Across policies, graph topologies, sensor profiles, and seeds, the worst-competitor score correlates more strongly with localization speed than an average-pair proxy in both 1D ($r=0.89$ versus $0.40$) and structured 2D sensing ($r=0.77$ versus $0.48$). Our results provide a practical way to certify and diagnose adaptive multi-agent sensing systems using the evidence they actually collect.

0.24

Fast Almost-Uniform Sampling of Random $k$-SAT Solutions

cs.DS cs.DMmath.PR ·2026-10-07 ·Kun He, Zhidan Li, Kuan Yang
theory
Abstract · A revised version with fast expected running time

We study approximately uniform sampling of satisfying assignments from random $k$-SAT formulas. Fix a constant $B>0$. For every $k\ge 2^{20}$ and density $0<α\le 2^k/k^{16}$, we prove that, with high probability over a random formula on $n$ variables, every accuracy $n^{-B}\le\varepsilon\le 1/2$ admits a sampler whose output distribution is within total variation distance $\varepsilon$ of the uniform distribution on satisfying assignments with expected running time $$\widetilde O_B\left(\mathrm{poly}(k,α+1)n^{2+η}\right),\quad 0<η\le 2^{16-k}.$$ In particular, the exponent of $n$ is $2+O(2^{-k})$ while the dependence on the width and density is polynomial with a universal degree. This improves the parameter-dependent polynomial exponent of the sampling algorithm of Chen, Lonkar, Wang, Yang, and Yin (STOC 2025) in the density regime $2^k/\mathrm{poly}(k)$. Our algorithm separates a high-degree core from the remaining variables. Every residual formula on the good variables is a deletion subformula of a common bounded-dependency formula. Adapting the rejection--repair framework of Liu, Wang, Yin, Zhang and Zhou (arXiv 2026), we construct an executable sampler for all such residual formulas using ordinary connected violation sets replacing weighted 2-tree witnesses and use it to simulate approximate block heat-bath updates on the core. To analyze the resulting core chain, we integrate out the good variables, represent the core marginal by an absolutely convergent polymer expansion and combine a component-swap variance comparison with an adaptive exploration bound. Absorbing single-component interactions into the local weights leaves interactions of cluster size at least two; retaining their additional decay gives a spectral gap of at least $1/(2s n^{2^{12-k}})$ where $s$ counts core components. Together with the residual sampler, this yields the stated running time.

0.23

The Silhouette Operator: Identifiability of Low-Rank Measures from One-Dimensional Projections

cs.IT math.STcs.LGmath.FA+1 ·2026-10-07 ·Robert A. Vandermeulen
Abstract

Structured recovery phenomena, such as restricted isometry properties in compressed sensing, have shown that high-dimensional objects can often be reconstructed from remarkably low-dimensional linear measurements. This work develops an analogous recovery framework for low-rank signed measures on $\mathbb{R}^2$, defined here as measures that can be expressed as finite sums of product measures with one-dimensional factors. The framework is based on linear operators, termed "silhouette operators," that map a measure to a fixed finite collection of one-dimensional linear pushforwards. The main results show that a suitably chosen collection of $2k$ projected marginals suffices to identify every compactly supported rank-$\le k$ signed measure, that this number is optimal, and that the projection directions cannot be chosen arbitrarily. The framework is also extended to higher-dimensional sums of product measures by establishing sufficient conditions under which collections of pairwise marginals identify the full model. Building on this framework, a computationally efficient estimator, termed "silhouette mixture estimation" (SME), is introduced for constructing a low-rank empirical measure from data by matching its one-dimensional projected marginals to the corresponding empirical marginals in Wasserstein distance. When combined with one-dimensional density estimators, SME yields an efficient nonparametric density estimator that performs strongly relative to a range of parametric, nonparametric, and deep-learning baselines in settings of moderate dimension and sample size.

0.22

Mixing Times of Switch Chains via High-Dimensional Expansion

cs.DM math.PRmath.CO ·2026-10-07 ·Sawyer Jack Robertson
Abstract · 52 pages, 3 figures

The switch chain is a Markov chain defined on the set of labelled realizations of a given graphical degree sequence. At each step, a pair of vertex-disjoint edges is chosen at random and the process attempts to replace them with a uniformly chosen perfect matching of the same four vertices, rejecting any proposal that would create a multiple edge. The resulting process is reversible with respect to the uniform distribution on all realizations. We investigate the mixing time of this chain by viewing realizations as the facets of a simplicial complex and studying a variant of the original process called the simplicial switch chain, which we analyze using tools from the theory of high-dimensional expansion. Our technical contributions include a proof that links of faces of sufficiently high codimension are strong spectral expanders and a comparison between the Dirichlet energies of large block updates and two-edge updates. Our main result is an $O(Δ^{2}m\log m)$ bound on the mixing time of both simplicial and classical switch chains whenever $m\ge CΔ^{8}$, where $m$ is the number of edges, $Δ$ is the maximum prescribed degree, and $C>0$ is an absolute constant. For sequences on $n$ vertices with fixed maximum degree, this proves that the chain mixes in $O(n\log n)$ steps, resolving a longstanding conjecture of Cooper, Dyer, and Greenhill and extending its conclusion to irregular degree sequences.

0.22

Improving the Constant in the Aharonov--Regev Theorem

cs.CC ·2026-10-07 ·Shuhong Gao
theory
Abstract

The Aharonov-Regev proof that $\mathrm{GapCVP}_{c\sqrt n}$ lies in $\mathsf{NP}\cap \mathsf{coNP}$ uses a verifier that tests dual lattice vectors; its proof gives $c=100$. We study this verifier through a linear program in which a certificate is modeled as a sample from a probability distribution on the dual lattice. This viewpoint gives an exact analysis of the close-target tests: adding gradient and Hessian checks increases the certified close radius from $σ^{-1}/4$ to $σ^{-1}\sqrt{3/16}$, and higher derivatives cannot improve it within this framework. With Gaussian certificates for far targets, we obtain $$\mathrm{GapCVP}_{c\sqrt n},\ \mathrm{GapSVP}_{c\sqrt n}\in \mathsf{NP} \cap\mathsf{coNP}$$ for every $c>2/(π\sqrt3)\approx0.3676$. An appendix gives a finite-bit implementation of the verifier. We also show that the Gaussian certificate is not always optimal and solve the unbounded-support program exactly for targets of order two, where its value is determined by the spectral capacity of an odd dual coset.

0.21

QoS-Aware Joint Concurrency and HARQ-Limit Design for HARQ-CC-Aided Slow Fluid Antenna Multiple Access

cs.IT ·2026-10-07 ·Sixu Han, Kai-Kit Wong, Hanjiang Hong, Haoyu Liang
theory
Abstract

Hybrid automatic repeat request with Chase combining (HARQ-CC) improves the reliability of slow fluid antenna multiple access (sFAMA), but its transmission limit must be jointly optimized with user concurrency. This paper proposes a quality-of-service (QoS)-aware joint design of concurrency level U and HARQ limit C for downlink HARQ-CC-aided sFAMA. Based on the selected-port signal-to-interference ratio (SIR) distribution, we derive multi-round decoding failure probabilities and develop a two-stage optimization that maximizes feasible concurrency and then selects the minimum C satisfying outage and mean-service constraints. We prove that the first feasible C minimizes service duration for a given U, and that feasible U values form an initial segment, enabling optimal bisection search. Numerical results agree with Monte Carlo simulations under full Jakes spatial covariance. The proposed adaptive FAS design supports six concurrent users, compared with three to five for fixed-HARQ FAS and two for HARQ-adaptive fixed-position-antenna (FPA), demonstrating that joint concurrency and HARQ adaptation expands the QoS-feasible region of FAS.

0.21

Spectral Efficiency Analysis of Massive MIMO-OFDM Systems with Double Quantization

cs.IT ·2026-10-07 ·Bengu Bilgic Keskin, A. Oguz Kislal, Ali Gorcin, Özlem Tuğfe Demir
Abstract

Massive multiple-input multiple-output (MIMO) achieves a high spectral efficiency (SE) by employing a large number of receive antennas in the uplink. Employing many receive antennas requires a large number of analog-to-digital converters (ADCs) and a high fronthaul data rate. Low-resolution ADCs can reduce power consumption and cost, while fronthaul quantization can reduce the required fronthaul data rate, at the expense of quantization distortion at both stages. In wideband orthogonal frequency-division multiplexing systems, ADC quantization is performed in the time domain, while fronthaul quantization may be applied after the discrete Fourier transform to transmit only active subcarriers. In this study, we investigate the achievable uplink SE over fronthaul links under double quantization and imperfect channel estimation. We derive a linear minimum mean-squared error channel estimator from the double-quantized pilot observations and obtain an achievable SE using the use-and-then-forget bound. Numerical results demonstrate that the ADC and fronthaul quantization resolutions have comparable impacts on the achievable SE. When the two resolutions differ, the lower resolution becomes the dominant factor limiting the SE. Furthermore, our numerical results show that six-bit ADC and fronthaul quantization achieve more than 90% of the SE attained in the ideal case.

0.21

Modeling and Performance Evaluation of Distributed Satellite Antennas: Space Is Much Closer Than One Might Think

cs.IT ·2026-10-07 ·Giacomo Bacci, Riccardo De Gaudenzi, Marco Luise, Luca Sanguinetti
Abstract · currently submitted to IEEE Transactions on Aerospace and Electronics Systems (14 figures, 1 table)

Distributed satellites can deliver broadband connectivity directly to commercial handsets. In the formation of arrays (FoA) architecture, cooperating satellites synthesize a large aperture whoseextent depends on their number, arrangement, and spacing, with intersatellite distances ranging from tens of meters to hundreds of kilometers. For representative S-band geostationary and low-Earth-orbit configurations, the Fraunhofer distance may exceed the nadir link range at spacings on the order of a hundred meters. Therefore, accurate assessment of FoA performance across satellite populations and spacings requires spherical-wave modeling for users in the radiative near field, below the formation's Fraunhofer distance. Planar models of satellite positions, array orientations, and user locations must also be assessed against spherical geometries accounting for orbital and Earth curvature. This paper combines these models to quantify how simplifying assumptions affect throughput and outage probability. Numerical results show that multi-antenna precoding based on a local plane-wave approximation within each satellite array closely matches the spherical-wave reference. A global plane-wave approximation can instead cause severe losses, including full outage. In the investigated scenarios, formation curvature has little impact, whereas neglecting Earth curvature overestimates low-Earth-orbit throughput. These findings clarify when geometric and propagation approximations remain reliable and demonstrate that, for large distributed antenna apertures, space is much closer than one might think.

0.21

Obstructions to $k$-colouring $H$-free graphs

cs.DM math.CO ·2026-10-07 ·Iain Beaton, Ben Cameron, Adam van Omme
Abstract

A graph is $H$-free if it has no induced subgraph isomorphic to $H$. In 2020, Chudnovsky, Goedgebeur, Schaudt, and Zhong characterized all graphs $H$ such that there are only finitely many minimal obstructions to $3$-colouring $H$-free graphs. In general, the minimal obstructions to $k$-colouring $H$-free graphs are the $(k+1)$-vertex-critical $H$-free graphs, those are, the $H$-free graphs $G$ with $χ(G)=k+1$ but $χ(G-v)=k$ for every vertex in $G$. In this paper we complete the characterization for all $k > 4$ by showing that there are onky finitely $k$-vertex-critical $H$-free graphs if and only if $H$ is an induced subgraph of $P_4+\ell P_1$ for some $\ell \geq 0$.

0.20

An $n^{0.3+\varepsilon}$-Approximation for Steiner $k$-Forest

cs.DS ·2026-10-07 ·Eden Chlamtáč
Abstract

We give an $n^{0.3+\varepsilon}$-approximation algorithm for the Steiner $k$-Forest problem, for any constant $\varepsilon>0$. As a function of $n$, this improves over the $O(\min\{\sqrt{n},\sqrt{k}\})$-approximation of Gupta et al. [ESA'07, TALG'10] which has stood for nearly two decades for the general case, as well as the later $n^{0.448}$-approximation of Dinitz et. al [APPROX-RANDOM'14, TALG'17] for the uniform weight case. On the other hand, we show that, due to lower bounds on the Densest $k$-Subgraph problem, the $O(\sqrt k)$-approximation for Steiner $k$-Forest likely cannot be improved. Specifically, we show that for any sufficiently small $\varepsilon>0$, an $O(k^{1/2-\varepsilon})$-approximation for Steiner $k$-Forest would surpass known degree-$n^{Ω(\varepsilon^2)}$ Sum-of-Squares integrality gaps for Densest k-Subgraph and refute the corresponding dense-versus-random conjecture.

0.20

Settling the Sample Complexity of Rényi Entropy Estimation

cs.IT cs.DSmath.ST ·2026-10-07 ·Qisheng Wang
theory
Abstract · 23 pages, 1 table

Rényi entropy estimation has been comprehensively investigated by Acharya, Orlitsky, Suresh and Tyagi (SODA 2015; IEEE Trans. Inf. Theory 2017) and consequent works, whereas only the sample complexity of Rényi entropy estimation of integer order has been settled. In this paper, we settle the sample complexity of Rényi entropy estimation of noninteger order, thereby completing the complexity picture of Rényi entropy estimation. Specifically, we show that for any noninteger $α> 0$, it is sufficient and necessary to use \[ Θ\!\left(\frac{d^{\max\{1/α,1\}}}{\varepsilon^{1/α}\log(d)} + \frac{d^{|1-1/α|}}{\varepsilon^2}\right) \] samples to estimate the Rényi entropy of order $α$ of an unknown discrete distribution over an alphabet of size $d$ to within additive error $\varepsilon$. For the upper bound, we reduce the bias using a refined polynomial approximation estimator for large probabilities. For the lower bound, we employ a different hard instance equipped with a new moment matching construction. The constructive moment matching has constant bounded high-order moments, while attaining a fixed ratio between the $α$-th moments, which is of independent interest.

0.20

On the Impossibility of Estimating the Average Mixing Time

cs.IT math.ST ·2026-10-07 ·Nathan A. Judd, Amedeo Roberto Esposito
Abstract

We establish nearly-matching converse bounds on the sample complexity of estimating the average mixing time of a Markov chain from a single stationary trajectory. Unlike the worst-case mixing time, which maximises the total-variation distance to stationarity over all initial states, the average mixing time averages this distance under the stationary distribution and can therefore be substantially smaller. Our converse bounds establish when estimation is impossible, and match the key dependencies in existing achievability bounds, up to logarithmic factors. Together, these bounds provide nearly-tight PAC guarantees for estimating the average mixing time.

0.18

The Hellinger Conjecture for Antipodal Boolean Functions

cs.IT math.COmath.PR ·2026-10-07 ·Vu Khac Ky
theory
Abstract · 15 pages

The balanced Hellinger conjecture compares the overlap of two output laws induced by a Boolean decision under binary symmetric noise. We prove that a single input coordinate minimizes this overlap among antipodal decisions, whose values reverse when all input signs are reversed, throughout the noise interval. The proof uses a normalized nuclear norm of the cross-class block of the Gram matrix of square-root channel densities. Multiplication by an antipodal decision exchanges the even and odd Fourier spaces preserved by the Gram matrix. A coordinate exchange attains the minimum cost of matching their spectra. A matrix contraction extends the sharp bound to arbitrary independent binary-input channels. Fixed coordinate flips and randomized decisions follow from the same argument and concavity. These extensions give bounds for information and testing, including linear Gaussian observations, and a weighted edge-boundary inequality.

0.18

Sequential Random Sampling PIR with Multiple Colluding Servers in DNA-Based Data Storage

cs.IT ·2026-10-07 ·Chen Wang, Natalia Silberstein, Eitan Yaakobi
Abstract

As DNA-based data storage evolves, protecting user privacy during data retrieval has become increasingly important. We study sequential random sampling DNA private information retrieval (SRS DNA PIR) with multiple colluding random sampling servers, where the database is partitioned into servers of equal size. We investigate the tradeoff between the download cost, defined as the expected number of queries, and the privacy leakage, measured by mutual information. We derive lower bounds on this tradeoff, including a bound given by an optimization problem. This bound is tight when each server stores two files, and we construct schemes that attain it. For servers of any size, we construct schemes that apply a single-server scheme to a randomly selected subset of servers.

0.16

Grover Search for a Threshold-Fraction Variant of the Travelling Salesman Problem: A Circuit-Level Construction

cs.DS quant-ph ·2026-10-07 ·Aayra Dhindsa Sawkmie
Abstract · 12 pages, 2 figures

The Travelling Salesman Problem (TSP) is a classic combinatorial optimisation problem which is NP-hard in optimisation form and NP-complete as a decision problem. The standard question asks for the shortest tour that visits every city exactly once and returns to the origin. This paper studies a threshold-fraction variant, which we call the robust TSP, formulated as a promise problem: given a distance threshold $K$ and a promised fraction $p$, decide whether at least a fraction $p$ of all tours have total length at most $K$, or whether no such tour exists. We show that Grover's quantum search algorithm decides this problem using $O(1/\sqrt{p})$ oracle queries, independent of the number of cities $N$, with each query implemented by a quantum circuit of polynomial size in $N$. For any fixed robustness level $p$, the total runtime is therefore polynomial in $N$. Measured in oracle queries, the quantum algorithm offers a quadratic improvement in the dependence on $1/p$ over classical random sampling, which needs $Θ(1/p)$ samples; because each quantum query is costly, however, it is faster in total running time only when $p$ is very small (roughly below $N^{-4}$), where the number of queries is no longer constant. We describe the problem formulation, detail the oracle construction (edge-length loading, sequential quantum addition, and threshold comparison), and analyse the query complexity to achieve the required result.

0.15

Envy-free Allocations with Individual Payments

cs.GT cs.DS ·2026-10-07 ·Robert Bredereck, Eva Deltl, Tanmay Inamdar, Pallavi Jain, Pranjal Pandey
survey
Abstract

When an envy-free allocation of indivisible goods does not exist, monetary transfers can restore envy-freeness. Existing work on fair division with subsidies, however, typically assumes that these payments are provided by an external source, an assumption that may be unrealistic in many applications. We address this limitation by allowing only monetary transfers between agents, with each agent's payments constrained by their individual budget. We show that while it is polynomial-time tractable to determine whether a given allocation can be made envy-free by payments under individual budgets, the general problem of computing such an allocation from scratch is NP-hard, even when agents have relatively large budgets. Motivated by this intractability, we conduct a parameterized complexity analysis, establishing fixed-parameter tractability with respect to the number of goods or to the joint parameter number of agents and good types, and we provide efficient algorithms in several special cases. For explicitly listed items, our type-based algorithm answers the envy-freeness part of an open question of T. T. Nguyen and J. Rothe, "Complexity Results and Exact Algorithms for Fair Division of Indivisible Items: A Survey."

0.15

Variational Bayesian Inference Based Progressive Multi-Target MIMO Sensing with Nuisance Parameters

cs.IT eess.SP ·2026-10-07 ·Jiayi Yao, Shuowen Zhang
Abstract · Submitted for possible publication

This paper proposes a progressive Bayesian multiple-input multiple-output (MIMO) sensing framework for multiple targets over multiple stages. We consider a practical yet challenging scenario where the targets' angles are unknown and random parameters to be estimated, while the targets' reflection coefficients are unknown nuisance parameters. With an initial prior probability density function (PDF) for the targets' angles, we progressively update the prior PDF for each sensing stage as the posterior PDF obtained from the previous stage, based on which Bayesian transmit beamforming optimization is performed to minimize the sum posterior Cramér-Rao bound (PCRB) in estimating the targets' angles and Bayesian sensing is performed with the help of new observations in this stage. To analytically characterize the intractable and high-dimensional posterior PDF with low complexity, we propose a variational Bayesian inference based approach which derives a surrogate posterior PDF in closed form with only polynomial complexity over the number of targets, in sharp contrast to existing numerical calculation approaches with exponential complexity. Numerical results validate the efficacy of our proposed framework in progressively refining sensing performance.

0.12

Polynomial Kernels for Interval Completion

cs.DS cs.CC ·2026-10-07 ·Zimo Sheng, Tian Bai, Mingyu xiao
Abstract

An interval graph is the intersection graph of a family of intervals on the real line. \textsc{Interval Completion} asks whether a given graph can be transformed into an interval graph by adding at most $k$ edges. Although the problem is fixed-parameter tractable when parameterized by $k$, whether it admits a polynomial kernel has long been an open question. We resolve this question by giving the first polynomial kernel for \textsc{Interval Completion}. Our main contribution is a parameter-preserving polynomial-time reduction from \textsc{Interval Completion} to \textsc{Odd Cycle Transversal} (OCT). Combining this reduction with the known randomized and deterministic polynomial kernels for OCT and a polynomial-time reduction back to \textsc{Interval Completion}, we obtain a randomized kernel with $\widetilde O(k^{18})$ vertices and $\widetilde O(k^{36})$ edges, and a deterministic kernel with $O(k^{36})$ vertices and $O(k^{72})$ edges. Here, $\widetilde O$ suppresses polylogarithmic factors in $k$.

0.10

Perfect Italian Domination on and Near Split Graphs: Algorithms, Hardness, and Approximation

cs.DM ·2026-10-07 ·Anand Babu N B, Ashwin Jacob, Manjusha M S, Renjith P
Abstract · Submitted to CALDAM 2027

A perfect Italian dominating function assigns a label from $\{0,1,2\}$ to each vertex so that the labels in the neighborhood of every zero-labeled vertex sum to exactly two. Although the associated decision problem (PID) is NP-complete on chordal graphs, the complexity status on split graphs remains open. We give an $O(n^4)$ time algorithm for split graphs on $n$ vertices. We note that this is in contrast to Roman, Italian, and perfect Roman domination, which are NP-complete on split graphs. We further show that this tractability boundary is tight: PID is NP-hard on graphs at deletion distance one from split graphs. We also provide dichotomy results for PID on restricted extensions of split graphs formed by attaching pendant stars, and establish inapproximability results for the same.

0.09

Attention via Black-Box Vector Search

cs.DS cs.LG ·2026-10-07 ·Stepan Zharkov, Krish Singal, Ashwin Padaki, Alexandr Andoni
Abstract · added funding acknowledgement

Sparse attention mechanisms estimate attention over $n$ tokens using a small subset of keys. Many existing approaches use maximum inner product search (MIPS) to retrieve the heaviest keys, which motivates the following question: given black-box access to a MIPS oracle, how many keys must be retrieved to output an $\varepsilon$-accurate attention estimate? We answer this question by unifying prior approaches through the framework of priority sampling. With a single MIPS index, we show that $Θ(\sqrt{n}/\varepsilon)$ retrieved keys are both sufficient and necessary. With $Θ(\log n)$ indices, we give an algorithm that retrieves only $O(\log n+1/\varepsilon^2)$ keys and prove that this is near-optimal. More generally, we design algorithms that establish a smooth tradeoff between the number of MIPS indices and number of retrieved keys. We then show that if we allow augmentation of keys and queries, we can bypass the above lower bounds: there exists a simple priority-sampling estimator using a single MIPS index and $O(1/\varepsilon^2)$ retrieved keys. When integrated into LLM inference, our algorithms outperform top-$k$ and sampling approaches used in prior work and yield attention approximation that scales favorably to long contexts.

0.07

Quantum Approximation Complexity of Classical Optimization Problems

cs.CC quant-ph ·2026-10-07 ·Stuart Hadfield
benchmark
Abstract

Classical approximation complexity asks what solution quality can be guaranteed with polynomial-time computation. The classes APX, PTAS, and FPTAS distinguish a fixed approximation ratio, approximation to any fixed accuracy, and approximation schemes whose running time is also polynomial in inverse accuracy. Their randomized counterparts are R-APX, R-PTAS, and R-FPTAS. We define bounded-error quantum counterparts BQ-APX, BQ-PTAS, and BQ-FPTAS. Membership requires a uniform quantum algorithm that, on every input, returns a feasible classical solution achieving at least the claimed approximation ratio with probability at least 2/3. Scores (objective values) must be efficiently classically computable. Running time includes parameter selection, preparation, measurement, decoding, and repetition. Many quantum optimization methods are used heuristically, and high benchmark scores alone do not establish these guarantees. We further establish a conditional hierarchy for logarithmic, polynomial, and exponential approximation factors. Assuming NP $\nsubseteq$ BQP, the quantum classes form a strict hierarchy. Problems based on prime factorization and discrete logarithms give conditional quantum-classical separations. Certified Maximum Order has an exact quantum algorithm, while any randomized polynomial-time algorithm guaranteeing at least an inverse-polynomial approximation ratio on every input would yield efficient factoring. Discrete-Logarithm Fitting has an exact quantum algorithm and a deterministic one-half approximation, but any fixed improvement over one half would give a randomized polynomial-time algorithm for the safe-prime discrete logarithm problem. Our results show, under explicit complexity assumptions, that quantum computation can improve worst-case approximation guarantees. A quantum-classical gap for common problems such as MaxCut or MaxSAT remains open.

0.05

Robotics, Control & Signals 122

ClimbLab: MATLAB Simulation Platform for Legged Climbing Robotics

cs.RO ·2026-10-07 ·Kentaro Uno, Warley F. R. Ribeiro, Yusuke Koizumi, Keigo Haji, Koki Kurihara +2 more
publishedhighly cited
Abstract · Author's version of a manuscript in the proceedings of CLAWAR 2021. The final published version is available at: https://doi.org/10.1007/978…

This paper presents an open-sourced MATLAB simulation and analysis platform dedicated to legged climbing robots. This simulator enables the design of any limbed robotic system as an articulated multi-body with a floating base and simulates it walking and climbing in an arbitrary environment. The main variable environmental parameters are inclination, gravity, and ground stiffness, and any point cloud can be installed as the terrain map. Furthermore, the simulator employs a rigid body dynamics engine. This paper first describes the simulator structure, and the computational flow and next presents the representative simulation examples where quadrupedal robots assumed gripping on the wall or climbing on the steep slope.

0.67

Constructive Safety-Critical Control for a Class of Underactuated Systems: A Hierarchical Approach

eess.SY math.OC ·2026-10-07 ·Massimiliano de Sa, Aaron D. Ames
published
Abstract · 8 pages, 2 figures. To appear at the IEEE Conference on Decision and Control (CDC), 2026

Underactuated systems with nontrivial geometry are abundant in practical robotic problems. However, little is known about constructive safety-critical control design in the presence of underactuation. In this work, we provide a constructive framework for synthesizing safe control architectures and control barrier functions for a class of quadrotor-like underactuated systems. By studying the geometry of their underactuation, we characterize when these systems have natural layered architectures useful for safety-critical control design. Using these architectures, we devise a constructive, hierarchical CBF synthesis procedure and provide conditions under which it succeeds.

0.64

RFPO: Rectified Flow Policy Optimization for Embodied Control

cs.RO ·2026-10-07 ·Ting Huang, Lisiyu Pan, Haoyu Wang, Zeyu Zhang, Siyuan Qian +4 more
code/data
Abstract

Flow-based policies provide an expressive framework for continuous robot control, but their iterative ODE integration incurs substantial inference cost. Naively reducing the integration budget can severely degrade control, since policies optimized under full-step execution are not explicitly constrained to remain reliable under coarse numerical integration. We refer to this mismatch as the few-step discretization gap. To address this problem, we introduce RFPO, a flow-policy optimization framework for reliable few-step execution. Reward-aware online Reflow rectifies student-induced transport paths during on-policy learning, making the resulting policy more robust to coarse integration. A frozen Gaussian PPO controller supplies complementary action-space supervision at full and intermediate integration budgets, while the deployed policy remains a single flow student executed with one Euler step. Across Unitree Go2, Boston Dynamics Spot, Unitree H1, and Unitree G1, RFPO consistently preserves full-step control performance under one-step execution, with one-step returns remaining within 2.4% of their corresponding 64-step values across both zero and random initialization. On Unitree Go2, one-step execution retains 98.5% of the 64-step reward while reducing onboard mean inference latency from 4.39 ms to 0.08 ms, yielding a 54.9x speedup. Real-robot experiments further validate stable one-step locomotion. Code: https://github.com/AIGeeksGroup/RFPO. Website: https://aigeeksgroup.github.io/RFPO.

0.64

Agentic RSR: Real-to-Sim-to-Real through Scene Reconstruction and Execution-Grounded Robot Policies

cs.RO cs.CV ·2026-10-07 ·Yihan Li, Yating Feng, Shengjiu Sun, Jianing Chen, Hao Ren +5 more
code/data
Abstract · 26 pages. Updated Real-to-Sim experiments and added model evaluations

A simulation of a real robot workspace must preserve task-relevant interactions, while policies developed in it must operate on observations available to the real robot. Yet scene reconstruction and policy development are often treated separately. We present Agentic Real-to-Sim-to-Real (Agentic RSR), a framework that links scene reconstruction, policy development, and real-robot execution through the same manipulation task. Given a workspace video, a task description, and a known robot model, an agent recovers metric scale, iteratively refines the scene using visual feedback, and checks task-relevant interactions in MuJoCo. A coding agent then develops an executable policy, progressing from privileged object poses to visual observations and randomized simulation. The policy can interleave multiple observations and actions within one invocation, while the agent uses execution feedback to continue, retry, or revise its approach. A shared task-level interface carries the policy and accumulated experience to the real robot, where fresh observations and safety checks guide execution. Using Claude Opus 5.5 through DeepSeek Harness (DSH), all 18 scenes across two robots pass the joint visual check; mean four-view Depth MAE against reference depth estimates is 0.0808 m, mean Lab $ΔE_{76}$ is 8.01, and mean grayscale SSIM is 0.7338. In real-robot experiments, the aggregate task success rate reaches 80% of the simulation task success rate, indicating substantial retention of simulated performance on hardware. Code and reconstructed scene data will be made publicly available.

0.61

NavGPT-3: Harnessing Context in a Hierarchical Navigation Runtime

cs.RO cs.AIcs.CLcs.CV+1 ·2026-10-07 ·Gengze Zhou, Yicong Hong, Jiazhao Zhang, Xunyi Zhao, Jian Zhou +7 more
code/data
Abstract · 36 pages, 14 figures. Project page: https://metacognitionai.github.io/NavGPT3/

Language models trained with long-horizon agentic reinforcement learning can generalize knowledge through reasoning, express precise actions, and pursue goals over many steps, raising the ceiling on what an embodied agent can understand and decide. Physical interaction, however, remains the domain of action policies, which provide dense, low-latency control. We present NavGPT-3, a harness that connects the two models, with an OS-like runtime built above it: reasoning, acting, and monitoring run as threads with their own context, tools, and permissions, while the runtime schedules them and decides which thread controls the robot's motion, so that the robot can react to sudden real-world events through interruption and thread switching. Beneath it, our action policy NavGPT VLA, trained on 19.28M examples, allocates visual tokens using codec allocation, in proportion to scene change; its 8B model alone reaches 76.94 SR on R2R-CE and leads RxR-CE with 82.39 SR. With the complete harness, NavGPT-3 sets the state of the art on R2R-CE (81.51 SR) and, for the first time, brings an autonomous agent to human level: on RxR-CE it matches human followers in success (90.43 vs. 90.4 SR) and path fidelity (78.47 vs. 77.7 nDTW) at 1 min 22 s per episode, versus roughly 3 min for a human. We comprehensively ablate the harness design and the interaction between the two models, showing how tools and the action policy shape the path from language-model reasoning to physical control: when NavGPT VLA executes the route, the reasoning loop shortens and the system's minimum reaction time falls from 3-19 s per language-model decision to 0.5-1 s per action-policy step (1-2 Hz). These results show that designing this embodied interface is central to connecting frontier language-model intelligence with low-level physical control. We will release all models, code, and evaluation records.

0.58

MagCilia: A Compact Magnetociliary Tactile Sensor with 3D Force Sensing for Robotic Contact Perception and Grasping Feedback

cs.RO ·2026-10-07 ·Yu Feng, Hao Wu, Haotian Guo, Haoming Liu, William Su +5 more
highly cited
Abstract

Robotic grasping and surface exploration benefit from simultaneous measurement of normal and tangential forces and from surface information obtained through contact. Here, we present a compact magnetociliary tactile sensor (MagCilia) that combines a flexible magnetic-cilia structure with a Hall sensor for 3D force sensing. Quasi-static finite element analysis is used to investigate structural deformation and magnetic responses under multidirectional loading. To reconstruct forces from the coupled magnetic channels, we propose causal history fusion regression (CHFR), which combines current magnetic-field measurements with their recent changes. Five-fold cross-validation grouped by calibration record yields root-mean-square errors of 0.40, 0.57, and 0.69 N for Fx, Fy, and Fz, respectively, with corresponding coefficients of determination of 0.93, 0.90, and 0.92. Robotic experiments demonstrate tangential-force-guided gripper adjustment and multi-axis load monitoring under external perturbations. Frequency-domain features of the reconstructed forces distinguish six surface categories with 99.39% accuracy in three-fold cross-validation grouped by acquisition session. An online robotic demonstration additionally identifies all six tested surfaces. These results demonstrate 3D force reconstruction, grasping feedback, and surface recognition using a single compact tactile unit.

0.58

Rephrase Before You Act: Characterizing and Mitigating Language Sensitivity in Vision-Language-Action Models

cs.RO cs.CLcs.LG ·2026-10-07 ·Mikey Watts, Yuchen Cui
code/datahighly cited
Abstract · 9 pages, 8 figures, 3 tables. Project page: https://sttawm.github.io/rephrase-before-you-act

Vision-language-action models (VLAs) are strikingly sensitive to instruction phrasing and do not inherit the language robustness of the vision-language models they are built on. A one-word edit can move success by tens of points: $π_{0.5}$ turns on a LIBERO stove 100% of the time for "switch on the stove" and 2% for "switch on the hot plate", and a $π_0$ checkpoint finetuned with rephrase augmentation still shows swings of up to 61 points. We characterize this sensitivity with statistically tested single-edit swings and an oracle phrase search, which shows that phrasing alone nearly closes the 21-point gap between in-distribution and out-of-distribution tasks. We then reduce it without modifying the policy. Because the sensitivity is systematic, it can be expressed as explicit rules: we score many phrasings of a few training tasks, have a large language model distill the evidence into ten to twenty rephrasing rules, and at deployment rewrite each incoming instruction once under these rules. The rules improve the frozen $π_0$ by 16 to 27% relative on twelve held-out tasks across adversarial, VLM-generated, and human-generated phrasings, with gains concentrated on out-of-distribution tasks. The pipeline replicates on $π_{0.5}$ and LIBERO, lifting in-finetune success from 93.6% to 97.8%. The method requires no retraining and no per-step verification, and applies zero-shot to unseen tasks and instructions. Project website: https://sttawm.github.io/rephrase-before-you-act

0.57

Near-Field Beamforming Design for Multi-IRS Wireless Beam Routing

eess.SP ·2026-10-07 ·Tao Wang, Changsheng You, Changchuan Yin
published
Abstract · 5 pages, 3 figures, accepted by IEEE Communications Letters

This letter investigates efficient near-field beamforming designs for wireless communication systems aided by multiple extremely large-scale intelligent reflecting surfaces (Multi-XL-IRSs). Previous studies have mostly assumed far-field channel conditions, based on which angle-based beamforming is adopted for the base station (BS) and IRSs. However, this approach may lead to degraded performance in practical near-field scenarios. To address this issue, we consider near-field channel modeling and formulate an optimization problem aimed at maximizing the received power at the user equipment (UE) by jointly optimizing the beamforming at the BS and IRSs. Given the intractability under the beam training (BT)-based communication framework, the problem is revisited and reformulated. Subsequently, an efficient algorithm based on alternating optimization (AO) is proposed to solve the reformulated problem. Numerical results demonstrate that, in typical setups, our AO-based beamforming designs provide over 170% and 20% received power gains over conventional angle-based beamforming and beam focusing, respectively. Furthermore, the power scaling law derived under far-field conditions no longer holds in near-field scenarios.

0.56

Fast and Robust Teach-and-Repeat Navigation Using MixVPR Visual Place Recognition*

cs.RO ·2026-10-07 ·Václav Truhlařík, Tomáš Pivoňka, Libor Přeučil
published
Abstract · 6 pages, 5 figures, 2025 European Conference on Mobile Robots (ECMR)

Teach-and-repeat navigation systems employing advanced visual place recognition techniques for localization exhibit key attributes for long-term mobile robot navigation, such as the ability to operate in unstructured and dynamic environments. However, existing solutions based on deep-learning techniques are computationally demanding, limiting their applicability. This work introduces a novel and efficient teach-and-repeat system built on the modern visual place recognition method MixVPR. Real-world testing demonstrated its ability to operate both indoors and outdoors, achieving robustness and navigation precision comparable to other state-of-the-art systems. In addition, its lower hardware requirements make it suitable for a wide range of robotic platforms and practical applications.

0.54

RoboPace: Contact-Aware Time-Optimal Retiming for Action-Chunk Policies

cs.RO ·2026-10-07 ·Mimo Shirasaka, Takehiko Ohkawa, Takuya Okubo, Nicola Scianca, Tatsuya Matsushima +1 more
Abstract · Project: https://robopace.airoa.io/

Robot manipulation data collection has been shifting from teleoperation toward robot-free demonstrations, through interfaces such as the Universal Manipulation Interface (UMI) or directly from human hands. Vision-Language-Action (VLA) policies trained on such data inherit the demonstrator's timing. Yet human timing does not directly transfer to robots: compliant hands tolerate fast contact, whereas robots may overshoot due to actuator and tracking limitations; conversely, robots can move faster in free space. This motivates a unified approach that reconciles execution speed with contact safety. We present RoboPace, an online retiming layer that preserves the policy's geometric path while adapting its timing, respecting the target robot's kinematic and dynamic constraints. It adapts execution speed based on predicted contact, jointly accounting for contact-dependent speed limits and the robot's motion constraints. The method requires no policy retraining and operates in real time. Across three contact-rich tasks on a dual-arm robot, faster uniform execution and physical-limit-only retiming largely fail. RoboPace instead achieves higher overall success than slow uniform execution while completing four of five commands in approximately half the time, retaining the reliability of slow execution without its time cost.

0.52

Cross-Embodiment Robot Foundation World Models with Latent Actions

cs.RO ·2026-10-07 ·Huang Huang, Sriram Yenamandra, Arjun Majumdar, Elie Aljalbout, Tushar Nagarajan +7 more
benchmarkhighly cited
Abstract

The diversity of robot embodiments and action spaces makes it challenging to build robot world models that generalize across different embodiments. We introduce the Latent Action-Conditioned Robot World Model (LAC-WM), which operates within a learned unified latent action space shared across diverse embodiments. This unified action space improves the world model's performance when adapted to previously unseen robot embodiments. We compare LAC-WM with an Explicit Action-Conditioned World Model (EAC-WM), which conditions on explicit motion labels. Our results show that explicit action conditioning leads to disjoint action representations across embodiments, limiting downstream performance when adapting to new robots. We evaluate both models on dexterous manipulation tasks and a modified LIBERO benchmark. LAC-WM improves downstream performance over EAC-WM by up to 46.7% on dexterous manipulation and 11.7% on LIBERO. Crucially, the unified latent action space allows LAC-WM's downstream performance to scale positively with the number of embodiments used during pretraining. In contrast, the disjoint action space in EAC-WM leads to decreased performance as the number of pretraining embodiments increases. These results highlight the importance of a unified action space for efficient cross-embodiment learning, addressing a key challenge in robotics. Project website: https://lacwm.github.io/

0.51

Lifelong small-object navigation in changing object layouts: a benchmark and method

cs.RO cs.CV ·2026-10-07 ·Jiagan Huang, Zikun Zhou, Zijian Ni, Hongpeng Wang, Guangming Lu +2 more
benchmarkcode/data
Abstract

Household robots need to continually navigate to different objects in the same environment, many of which are small and portable, such as tools and toys. Their small visual footprint and frequent occlusion make reliable observation difficult, and they may be moved by people without the robot observing the changes. We formulate this challenging task as Lifelong Small-object Navigation in Changing Object Layouts (LiSoNav-COL). Agents must seek suitable viewpoints for reliable observation, accumulate and reuse scene knowledge to efficiently locate subsequent targets, and update outdated memory after object relocation. To eliminate the need for prior scene scanning, we also require agents to start navigation with empty scene memory. Although practical, this task still lacks benchmarks designed around its defining assumptions. To bridge this gap, we introduce LiSoNav-Eval, a dedicated benchmark spanning 28 indoor scenes with 45 small-object categories. Its lifelong navigation sequences include both unchanged and relocated targets to evaluate memory reuse and adaptation to object relocation. To address this challenging task, we propose a navigation method based on multi-view Inspection with Viewpoint-Anchored Memory, dubbed IVAM-Nav. IVAM-Nav actively observes supporting surfaces from complementary viewpoints for reliable small-object perception and anchors the resulting memory to their observation viewpoints, supporting relational memory reuse and revalidation under similar viewing conditions. Extensive experiments on LiSoNav-Eval demonstrate favorable performance of IVAM-Nav against representative methods. Benchmark analyses also show that smaller objects, larger environments, and longer relocation distances pose greater challenges. The dataset and code are available here.

0.51

MCFR: A Mask-Guided Coarse-to-Fine Regression Framework for Robust Multi-Variant Board-to-Board Connector Assembly

cs.RO ·2026-10-07 ·Guanghui Shen, Song Wang, Dan Wu
published
Abstract · 8 pages, 7 figures, 1 table. Accepted for presentation at the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS…

Automated insertion of board-to-board (BTB) connectors in 3C manufacturing requires both high visual accuracy and strong deployment robustness. This problem remains challenging because multi-variant connectors exhibit significant morphological and appearance variations, making stable cross-variant generalization difficult, while the mismatch between training and deployment under fixed-view inspection settings induces background spurious correlation and degrades real-world performance. To address these issues, this paper proposes MCFR, a Mask-Guided Coarse-to-Fine Regression framework for multi-variant BTB connector assembly. By introducing an object-aware mask prior and explicit photometric refinement, the proposed method suppresses background interference and improves alignment accuracy and robustness in practical deployment. Experiments on a self-constructed multi-variant dataset, a BTB batch insertion testbed, and a real smartphone assembly task show that MCFR consistently outperforms representative baselines and achieves an average real-world insertion success rate of 99.25%. These results demonstrate the effectiveness and practical potential of MCFR for automated assembly of multi-variant BTB connectors.

0.51

RoboRender: Robot-Oriented Video Generation for Visual Sim-to-Real Transfer

cs.RO ·2026-10-07 ·Huang Huang, Wensi Ai, Ziyu Chen, Youhui Wang, Zijian Du +4 more
Abstract

Simulation enables large-scale, low-cost robot data generation, but policies trained in simulation often fail to transfer to the real world due to the sim-to-real visual discrepancies. Existing approaches often rely on intermediate representations, which can discard rich semantic information or require additional perception modules at deployment. We address this visual sim-to-real gap with RoboRender, a framework that converts simulated trajectories into photorealistic RGB videos for policy learning. RoboRender trains a robot-oriented video generation model conditioned on simulated depth videos, language instructions, and robot RGB mask videos, preserving simulator geometry, robot motion, and action labels while synthesizing realistic textures, backgrounds, and distractors. The generated RGB videos are paired with simulator-provided states and actions to train policies for zero-shot real-world deployment. On robot video test sets, our video model outperforms depth-conditioned video generation baselines in generation quality. In real-world experiments across pick-and-place, articulated-object manipulation, and mobile manipulation tasks, policies trained on RoboRender-generated data achieve a 71% average success rate, outperforming raw simulation renderings and conventional visual domain randomization by approximately 7.1x and 3.6x, respectively. We further show that policy performance improves with more generated videos per simulation trajectory, increasing opening-task success by 65 percentage points. These results demonstrate that generative video rendering mitigates the visual sim-to-real gap for zero-shot policy transfer. Project website: https://robo-render.github.io/.

0.50

Adaptive Risk-Certified Event-Triggered Replanning for Dynamic Navigation

cs.RO eess.SY ·2026-10-07 ·Richie R. Suganda, Bin Hu
benchmark
Abstract · for associated mpeg file, see https://nail-uh.github.io/corl2026.github.io/

Safe navigation in dynamic environments requires robots to plan under obstacle predictions whose errors are uncertain, non-stationary, and can induce rare but safety-critical failures. Existing control-barrier-function safety filters can reject immediately unsafe controls, but they provide little guidance on when the current finite-horizon planning mode itself is becoming unsafe as prediction uncertainty evolves. We propose Conformal Event-Triggered Risk-Certified Replanning (\emph{CERT-Replan}), a framework that uses calibrated barrier risk as an early-warning signal for replanning. CERT-Replan calibrates horizon-indexed obstacle-prediction residuals online and uses the resulting uncertainty radii to evaluate dynamic-obstacle safety margins. A one-step safety filter protects the next applied control, while a horizon-level risk monitor evaluates the upper-tail CVaR of predicted barrier-violation losses along the current MPC rollout. When this risk exceeds an allocated budget, CERT-Replan rejects the current planning mode and selects a lower-risk alternative, such as a different speed profile, corridor, or homotopy class, rather than repeatedly correcting the same nominal plan. In a non-stationary benchmark, CERT-Replan achieves an \(83.3\%\) collision reduction relative to the safety-filter-only baseline \textcolor{black}{and a \(77.8\%\) reduction relative to simple replanning triggers}, while reducing average safety-filter intervention by \(32.4\%\). \textcolor{black}{With Trajectron++, CERT-Replan achieves \(96\%\) collision-free operation. Hardware experiments and onboard runtime profiling demonstrate computational feasibility.}

0.49

RobotWorld: Benchmarking Multimodal Agents for Robot Use Across Diverse Tasks and Embodiments

cs.RO cs.LG ·2026-10-07 ·Zhiqin Yang, Chenxin Li, Xiaomeng Hu, Yibin Liu, Weidong Huang +28 more
Abstract · 62 pages, 25 figures

General-purpose agents increasingly write code, use tools, and complete complex digital tasks, raising the question of how far these capabilities carry into the physical world. To investigate this, we introduce RobotWorld, a challenging simulation testbed for robot use: turning instructions and observations into physical task execution through robot interfaces. Its 84 tasks span manipulation, mobile manipulation, locomotion, driving, and aerial control, with explicit interaction budgets and executable success checks. By analysing task outcomes alongside execution traces, we identify both the capabilities that transfer and the gaps that prevent reliable completion. Furthermore, we find that current agents can construct sophisticated perception and control workflows, including image segmentation, camera calibration, spatial estimation, and dynamics-based computation. These capabilities, however, do not consistently compose into successful behaviour: agents lose task-relevant object states despite reaching commanded poses, fail to correct ineffective actions, recover too late, or mistake unfinished tasks for completion. This uneven transfer also differs across models: Astra succeeds more often on spatial and constrained-contact goals, whereas Opus 5.5 succeeds more often on continuous-balance and timed-interaction goals. By linking these outcomes to execution behaviour, RobotWorld provides both a rigorous proving ground and an empirical account of the remaining capability gaps, thereby establishing concrete targets for training and designing more reliable physical-world agents.

0.48

OpenViTac: Learning and Benchmarking Visuo-Tactile Policies in a Unified Sim-and-Real Framework

cs.RO ·2026-10-07 ·Yifan Wu, Qin Li, Nan Min, Guojin Zhong, Haoyu Zhao +12 more
benchmark
Abstract · Project website: https://fvl-repo.github.io/OpenViTac/

Tactile feedback provides embodied agents with physical information beyond visual observations, enabling more reliable interaction with the real world. However, despite the rapid progress of vision-tactile-language-action (VTLA) policies, there remains a lack of unified benchmarks for evaluating tactile-enabled robot manipulation across simulation and the real world. To address this gap, we introduce OpenViTac, a visuo-tactile manipulation benchmark for evaluating robot policies across simulation and the real world. OpenViTac organizes contact-rich manipulation into four tactile-relevant capability dimensions and provides paired simulation-real-world settings for consistent evaluation of VLA, WAM, and VTLA policies. Building upon this benchmark, we investigate how different tactile representations and integration strategies affect the performance of pretrained VLA models. Correspondingly, we introduce OpenVTLA, a tactile augmentation framework that combines the best-performing representation and integration strategy. Furthermore, we leverage the paired benchmark setting to study sim-real co-training and analyze factors affecting cross-domain policy learning. Together, OpenViTac provides a unified platform for evaluating and advancing visuo-tactile robot manipulation.

0.48

Self-Supervised Deconvolution of In-Air Sonar Images Using Sensor Ego-Motion

eess.SP ·2026-10-07 ·Jan Steckel
Abstract

Broadband in-air sonar sensors with a sparse microphone array form images through delay-and-sum beamforming, which leaves high sidelobes and an angular resolution that degrades strongly towards the sides of the field of view. Deconvolving these images requires access to the point-spread function of the complete imaging chain, including the emitter spectrum and directivity and the array imperfections. In practice, these aspects are hard to calibrate for. In this letter, we propose a self-supervised method that learns a deconvolution network together with a physics-structured model of the point-spread function, requiring only the sonar data itself and the odometry of a moving robot on which the sensor is mounted. Using acoustic flow warping, the internal representation of the model is warped and subsequently rendered through the learned point-spread function. This is then compared to the measured images under a speckle likelihood in accordance with the statistical model of image formation. In simulation, the learned model recovers the unknown emitter directivity, and the deconvolved images reach a constant angular resolution of 5 to 6 degrees over the full field of view, against 12 to 30 degrees for delay-and-sum and 7 to 23 degrees for delay-multiply-and-sum with coherence factor. The method outperforms classical deconvolution with the nominal point-spread function, and even outperforms CLEAN with the true one.

0.48

Do Better Visual Representations Always Lead to Better End-to-End Autonomous Driving?

cs.RO cs.CV ·2026-10-07 ·Zihao Zhang, Haochen Tian, Tianyu Li, Changhui Jing, Jingliang He +3 more
code/data
Abstract

Visual foundation models (VFMs) are increasingly integrated into end-to-end autonomous driving for their powerful representations, yet it remains unclear when these representations improve driving performance. To investigate this question, we introduce ViRA, a planner-agnostic visual representation alignment framework that keeps the planner architecture and inference cost unchanged. Our study reveals three findings: (1) VFM-guided visual representations consistently improve driving performance across diverse end-to-end planners, with gains extending to zero-shot closed-loop evaluation. (2) The choice of VFM target matters for planning performance, and alignment to a different VFM can further benefit planners with pre-trained VFM encoders. (3) Auxiliary perception supervision reduces sensitivity to VFM target selection, narrowing the EPDMS spread across five targets from 2.7 to 0.5 points and potentially compensating for less effective VFM targets. Guided by these findings, we develop ViRA-Diffusion, a diffusion-based planner trained without auxiliary perception supervision, which achieves 92.3 EPDMS on NAVSIM v2 navtest, outperforming recent methods in our comparison by at least 1.9 points. The results motivate jointly considering target selection and planner supervision when integrating VFMs into end-to-end autonomous driving. The results and demo are available at https://github.com/OpenDriveLab/ViRA.

0.48

OmniHOI: Dexterous Hand-Object Interaction from Monocular Human Video

cs.RO ·2026-10-07 ·Ting Mao, Yanming Shao, Ziheng Wang, Haoyu Liu, Yiqun Wang +2 more
code/data
Abstract · The project page is available at https://hydrofoooil.github.io/OmniHOI/

Monocular videos of human manipulation provide abundant dexterous demonstrations, yet reconstructing hand-object interaction from a single view and transferring it to robot hands remain difficult, limiting their direct use for robot execution. Prior methods either require task-specific RL training, limiting scalability, or assume clean motion-capture trajectories and thus cannot operate directly on video. We present OmniHOI, a pipeline that turns an RGB video of hand-object interaction into an interaction-faithful trajectory on dexterous hands. The key idea is to enforce physical consistency using the evidence available at each stage: image evidence during reconstruction, contact geometry during retargeting, and dynamics during physics-in-the-loop refinement. Each stage optimizes the corresponding representation directly, correcting errors before they propagate downstream or must be absorbed by a learned policy. Across 150 motion-capture trajectories transferred to each of five dexterous hands with 6 to 22 DoF, we achieve 39-89% success, compared with at most 31% for prior transfer methods. On 60 monocular video clips, we achieve 53% success, compared with 28% for the best prior video-to-robot pipeline. Its trajectories also execute on a real bimanual robot across diverse tasks.

0.47

Safety-Oriented Trajectory Ranking from BEV Perception Outputs

cs.RO cs.CV ·2026-10-07 ·Leon Pohl, Martin Frauendorfer, Corinna Schmitt, Mirko Maehlisch
publishedsurveycode/data
Abstract · Accepted at the 2026 IEEE International Conference on Vehicular Electronics and Safety (ICVES 2026). 6 pages, 4 figures, 3 tables. Code: htt…

Bird's-Eye-View (BEV) representations are increasingly used as compact inputs for autonomous driving perception and planning. However, lightweight and interpretable methods for evaluating candidate ego trajectories directly on BEV outputs remain useful for vehicular safety workflows, rapid prototyping, and human-in-the-loop validation. This paper presents a modular trajectory ranking pipeline that consumes BEVFusion object detections and map segmentations derived from nuScenes sensor data and ranks manually specified candidate trajectories according to safety-first criteria. The ranking prioritizes avoiding collisions and walkway violations, then smaller end distance to the target, and finally path efficiency. To study whether these rule-based rankings resemble human judgment, we created seven BEV traffic scenarios and compared the script output with preferences from a survey of 21 participants. Across 146 pairwise votes, 124 matched the trajectory ranked first by the evaluator, corresponding to an overall agreement of 84.9 %. The results indicate that simple BEV-centric safety heuristics can approximate human preferences in constrained trajectory choice tasks while remaining transparent enough for analysis, dataset curation, and autonomous driving studies.

0.46

Safe Control of Semi-Explicit Differential-Algebraic Systems: Control Barrier Functions with SOS Verification

eess.SY math.OC ·2026-10-07 ·Mohamad H. Kazma, Hongchao Zhang, Abdallah A. Albustami, Ben Wooding, Meiyi Ma +2 more
Abstract

Differential-algebraic equation (DAE) systems arise in power networks, multibody mechanics, and chemical processes, where algebraic constraints couple dynamic and algebraic states. Standard control barrier function (CBF) methods, designed for ordinary differential equations, fail for DAE systems because the algebraic constraints impose compatibility requirements that may render safety filters infeasible. This paper develops a safety-critical control framework for semi-explicit DAE systems using CBFs that consider compatibility with the algebraic constraints, termed DAE-aware CBFs. We construct projected vector fields on the constraint manifold to account for the hidden dynamics, while ensuring compatibility conditions that keep the algebraic constraints satisfied. We then present sum of squares (SOS) verification conditions that certify both correctness and feasibility of the resulting DAE-aware CBF. The framework is validated on power system case studies.

0.46

Predictive Beamforming via Echo-Beam Modality Separation with Cross-Attention for ISAC Systems

eess.SP ·2026-10-07 ·Seokjun Park, Jiwhan Jang, Jinseok Choi
published
Abstract

In this paper, we propose a deep learning-based predictive beamforming algorithm for integrated sensing and communications in vehicle-to-infrastructure networks, where echo signals are leveraged to enhance communication performance. We formulate a sum spectral efficiency maximization problem under a sensing signal-to-interference-plus-noise ratio (SINR) constraint and construct a penalty-based unconstrained surrogate objective for network training. Based on this formulation, the beamforming matrix is predicted using sensing echoes and historical beamformers from previous time slots. To this end, we develop a predictive beamforming network that processes sensing echoes and beamforming inputs through separate encoders to preserve their modality-specific characteristics. This separate design enables each modality to learn rich and complementary representations which are more effectively exploited through a cross-attention mechanism compared to naive input fusion. The resulting features are fused via cross-attention, while gated recurrent units capture temporal dependencies across time slots. Simulation results validate the performance of the proposed method.

0.45

Design optimization of tendon-driven robots considering tendon wrapping and shortcut

cs.RO ·2026-10-07 ·Yuta Sahara, Kento Kawaharazuka, Kei Okada
published
Abstract · Accepted to Advanced Robotics, website: https://haraduka.github.io/opt-wrap-shortcut

This study proposes a multi-objective optimization method for tendon-driven arm design, addressing the inherent trade-off between joint torque and arm thickness through tendon wrapping and shortcut effects. We formulate the problem to simultaneously maximize torque and minimize thickness, solved using NSGA-II. The algorithm optimizes tendon routing, pulley configurations, and attachment points. Results reveal Pareto-optimal designs with effective moment arm expansion via strategic shortcuts, providing practical guidelines for compact high-torque arms. The optimized configurations demonstrate non-trivial patterns beyond conventional design intuition, offering new insights for engineering efficient robotic systems.

0.43

Distributionally Robust Frequency Response Estimation Using Non-Steady-State Frequency-Domain Data

eess.SY ·2026-10-07 ·Stefan Damjanovic, John W. Simpson-Porco, Klaske van Heusden
theory
Abstract

This paper develops a distributionally robust optimization (DRO) framework for identifying the frequency response function (FRF) of an unknown linear time-invariant system from repeated noisy non-steady-state input-output measurements. Building on a recent frequency-domain extension of Willems' fundamental lemma, we formulate FRF estimation as a DRO problem, wherein robustness is desired with respect to the unknown probability distribution generating the noisy data. The proposed estimator computes FRF estimates by minimizing a convex upper bound on the DRO cost and allows for frequency-dependent calibration of the Wasserstein metric; the tractable upper bound consists of an empirical fit and a regularizer that penalizes the largest weighted sensitivity to spectral perturbations. Under suitable assumptions, we prove that the FRF estimation error is upper bounded by the optimal value of the DRO problem. The method is validated via a simple numerical study, which demonstrates improved performance under severely heavy-tailed noise.

0.43

Safe Learning of Adaptive Control Policies for Remote Patient Monitoring

eess.SY ·2026-10-07 ·Ramanan Tamizholi, Siddharth Chandak, Isha Thapa, Nicholas Bambos, David Scheinker
published
Abstract · Accepted at IEEE Healthcom 2026

Remote Patient Monitoring (RPM) enables continuous observation of patients in their daily environments, improving both health outcomes and quality of life. A key challenge in RPM is determining the optimal monitoring intensity, while balancing patient safety and monitoring costs. This problem is further complicated when system parameters, such as transition probabilities and costs, are initially unknown. We develop a learning-based control framework that estimates these parameters and adapts the monitoring policy in real time. The proposed approach is an online model-based reinforcement learning algorithm tailored to RPM, with patient safety explicitly prioritized during exploration. We provide theoretical guarantees on safety and convergence to the optimal policy. Simulation results show that the algorithm converges to the optimal threshold-based policy, maintains low treatment costs, and reduces the risk of patients reaching critical health states.

0.42

Second-order optimization of variable projection SVM models and road abnormality detection

eess.SP cs.LG ·2026-10-07 ·Andrea Angino, Matthias Voigt, Rolf Krause, Tamás Dózsa
publishedunusual
Abstract

We introduce a novel second-order optimization framework for minimizing so-called variable projection functionals. We demonstrate that the proposed framework is especially usefulfor the training of variable projection based kernel methods. In particular, the problem of efficiently training variable projection support vector machines (VP-SVMs) is considered. We show the effectiveness of the proposed training methodology in a real-world application, namely we demonstrate how second-order trust region algorithms can be used to train VPSVM models to recognize road surface abnormalities based on 1D signals obtained from a tire sensor.

0.40

Many Ways to Succeed: Diversity-Driven RL Fine-Tuning for VLA Generalization

cs.RO cs.LG ·2026-10-07 ·Haoru Li, Jinmei Liu, Zhiyong Wang, Xiaoming Li, Zhenhong Sun +3 more
Abstract

Reinforcement learning (RL) fine-tuning improves vision-language-action (VLA) policies through closed-loop experience, yet generalization beyond the fine-tuning distribution remains limited. Our analysis reveals a selective reshaping of exploration: RL contracts behavior globally, yet diversifies successful trajectories, elicits success with fewer rollouts, and covers more of the latent task-valid solution space than supervised fine-tuning. Broader successful-mode coverage may provide alternative strategies under distribution shifts. Inspired by this, we introduce DRIVE (Diversity-driven RL fIne-tuning for VLA gEneralization), which turns successful-behavior diversity into an explicit RL objective. DRIVE groups rollouts under matched task conditions, compares their trajectories with temporal alignment, and derives a success-conditioned intrinsic reward from relative behavioral diversity. This design encourages broader coverage of feasible solutions without rewarding diverse failures or superficial timing differences. Across LIBERO-Plus, ManiSkill3, and RoboTwin 2.0, DRIVE improves the average out-of-domain (OOD) performance over vanilla RL fine-tuning by 5.3 points on $π_0$ and 2.0 points on $π_{0.5}$. On a dual-arm AgileX PiPER-X platform, DRIVE further increases average OOD success from 64.1% to 73.3% (+9.2 points), demonstrating gains that persist under physical deployment.

0.40

CMP-IRRT*: A Perception-Assisted Height-Adaptive Planner for Quadruped Robots

cs.RO ·2026-10-07 ·Mingfan Zhao, Wendong Mao, Zhongfeng Wang
publishedcode/data
Abstract · Accepted at the 18th Asian Conference on Machine Learning (ACML 2026)

Quadruped robots can traverse low obstacles, but many 2D planning pipelines still model obstacles as binary occupied regions and rely on sampling-based search that can be inefficient under a limited budget. We propose a perception-assisted height-adaptive planning framework based on CMP-IRRT*, a Channel Mamba PointNet-guided Informed RRT* planner. Given a calibrated top-view RGB observation, the perception module estimates obstacle regions and converts depth predictions into a ground-relative height map. The planner then performs height-conditioned collision checking, treating high obstacles as blocked while allowing low obstacles to be traversed, and uses the CMP guide to bias sampling toward promising regions while retaining standard free-space and informed sampling fallbacks. Experiments on 2D planning benchmarks show that CMP-IRRT* reduces explored nodes and iterations compared with classical and neural-guided baselines, and a controlled ablation supports the contribution of the Mamba-based guide. In constructed traversability-aware scenarios, the proposed planner reduces path length by up to 16.3% when low obstacles are traversable, and a Unitree Go2 demonstration further shows executable bypassing and traversal behaviors. Our code is publicly available at https://github.com/MingfanZhao/height-adaptive-planner.

0.40

Long-WAM: Scaling the Context of World-Action Models

cs.RO cs.AIcs.CV ·2026-10-07 ·Wei Huang, Bohan Zhang, Chenzhi Liu, Isabella Liu, Shuai Yang +11 more
Abstract

Real-time robot control demands enough visual history to infer motion and task progress, but processing that history can delay action. We present Long-WAM, a model-system framework for scaling the context of causal world-action models under real-time control constraints. Our central finding is that access to history is not the same as using it: longer histories pay off far more when the video foundation is pretrained autoregressively (AR). We first learn causal prediction from robot and egocentric videos without action labels, then preserve this history-to-future structure during world-action adaptation. On RoboCasa GR-1, increasing context from 0.0 to 19.2 seconds raises success from 63.3% to 78.7%, whereas a bidirectionally pretrained initialization shows no net gain; robot-domain AR pretraining further raises peak success on GR-1 and LIBERO-Long. Long-WAM also achieves the best results among compared methods on LIBERO-Long, RoboTwin 2.0, and DOMINO. Streaming observation encoding, asynchronous execution, and hardware-specific acceleration enable deployment on RTX 5090, DGX Spark, and Jetson AGX Thor without dropping future prediction; on RTX 5090, each action chunk, including future-video latent prediction, takes 107.4 ms. Real-time deployment on Unitree G1 and YAM supports dynamic and long-horizon manipulation, including 95% success on dynamic cup stacking, where Pi0.5 and Fast-WAM succeed in none of 20 trials. As a memory-informed executor, Long-WAM also complements higher-level planning in composite tasks.

0.40

iAm.md: Robot Skill Self-Assessment through Agentic Introspection for Unknown Open-Vocabulary Domains

cs.RO cs.AI ·2026-10-07 ·Vincenzo Guarino, Emanuele Musumeci, Vincenzo Suriani, Daniele Nardi
publishedcode/data
Abstract · 7 pages, 2 figures, 1 table. Accepted at the 13th Italian Workshop on Artificial Intelligence and Robotics (AIRO 2026). Project page: https:…

Agentic AI based on Large Language Model generalization capabilities offers a wide range of potential applications, including planning for embodied tasks. For example, embodied agents based on Foundation models can generate plausible plans in autonomous robotics scenarios. Due to limited context windows or hallucinatory phenomena in the next-token prediction formulation, behaviors may be generated without establishing whether the deployed robot and the observed environment actually support the requested operation, in what we call a "grounding failure". Thanks to the recent improvements in reasoning capabilities of foundation models, autonomous robot behavior generation problem can be formulated as a code generation problem. We present iAm.md, a Markdown standard and generation framework, that allows anchoring this process in complementary forms of deployment evidence. Through open-vocabulary semantic mapping, we combine local vision-language detections and object segmentation and refer them to persistent object records in this intermediate standardized representation, allowing agentic introspection. We then study this new technique on a simulated TIAGo, on navigation-and-manipulation tasks, showing how this standardized representation jointly supports skill self-assessment and executable task generalization.

0.39

End-to-End Autonomous Generation of Human Assembly Plans

cs.RO cs.GR ·2026-10-07 ·Faustin Arion von Arx, Millicent Schlafly, Mark D. Fuge
code/data
Abstract · 22 pages, 10 figures

Turning a CAD design into an assembly plan is still largely done by hand, requiring engineers to reason about geometric feasibility, tool access, stability, and the ergonomics of human assembly. In this work, we encode long-established design for assembly (DfA) principles into a contained, end-to-end approach for generating assembly plans. Our approach takes only a mesh assembly and produces either a step-by-step assembly manual or a structured failure report, requiring no joint metadata, fastener annotations, or additional information. Four major components of a manufacturing plan are addressed autonomously: an assembly tool list, the assembly sequence and subassemblies, an assembly manual, and design feedback for improving assemblability. For determining the sequence plan, we systematically disassemble the object in a physics simulator and apply a cost function that encodes DfA principles. Manual generation, tool labelling, and assembly feedback rely primarily on multimodal large language models. Compared with a baseline that always removes the outermost part first from Tian et al., DfA-aware sequence planning reduces simulated assembly time, measured with a robot-arm assembly-time proxy, by 35% on 136 assemblies of 5 to 30 parts. The correct tool is selected for 88.6% of assembly steps. A vision-language model judge compares the generated manuals against ablated variants, identifying which page elements carry the information a reader needs. The presented approach and open-source code are available for use by engineers or AI agents looking to rapidly accelerate the creation of manufacturing plans for a given product design.

0.39

TMT: Runtime Backdoor Detection for Vision-Language-Action Policies on Unseen Tasks

cs.RO ·2026-10-07 ·Zirun Zhou, Jingfeng Zhang, HaoChuan Xu, Xizhe Zhang, Elliott Wen +2 more
code/data
Abstract

Backdoored vision-language-action (VLA) policies can preserve benign task performance while producing malicious actions when a trigger appears. Detecting such activation is difficult because malicious behavior can comprise individually plausible actions, while unfamiliar tasks introduce legitimate changes in observations and behavior. We introduce TMT, a runtime backdoor detector based on Token Manifold and latent Transition modeling. Trained on benign rollouts, its two branches assess input-token structure and prediction errors in adjacent-layer latent dynamics. A suspicious rollout identified by the token manifold branch, once confirmed through latent deviations, guides transition selection for subsequent monitoring. We further explore policy purification through self-distillation: a frozen copy of the backdoored policy provides benign-input actions to supervise a student on paired benign and triggered observations, without requiring a separate clean reference policy. For evaluation, we adapt traditional backdoor detectors and repurpose anomaly and failure detection methods as VLA backdoor detectors. In a post-hoc comparison with ten baselines, TMT achieves state-of-the-art backdoor detection performance on unseen tasks across three VLA backdoor attacks. Our project page is available at https://zzr42.github.io/tmt/.

0.39

Design of a Fully Actuated 4-DOF Robotic Finger With Joint-Specific Hybrid Remote Actuation

cs.RO ·2026-10-07 ·Hyojae Kang, Hyun-mok Jung, Joonho Lee, Dongil Park, Hyunmin Do +2 more
Abstract · 10 pages, 10 figures. This manuscript has been submitted for possible publication

This paper presents a fully actuated 4-DOF robotic finger using a joint-specific hybrid remote-actuation architecture. The metacarpophalangeal (MCP) joint is driven by two coordinated rigid-link transmission sets, whereas the proximal interphalangeal (PIP) and distal interphalangeal (DIP) joints are independently actuated by closed-loop wire transmissions incorporating circular rolling-contact joints (RCJs). A larger transmission radius is used at the PIP joint than at the DIP joint. The RCJ wire geometry maintains the total wire-loop length during joint rotation, and the DIP wire routing is designed so that PIP motion does not affect its differential actuation for a fixed MCP configuration. The distal wire transmission, MCP linkage, and fingertip kinematics are analytically modeled. In experiments, the transmission behavior was quantitatively evaluated from the ball-screw displacements measured using ArUco-marker tracking. MCP actuation produced measurable displacements of the PIP and DIP transmission units, whereas isolated PIP and DIP actuation with the MCP fixed supported the intended mechanical decoupling between the distal transmissions. The mean peak fingertip forces under isolated MCP, PIP, and DIP actuation were 21.28 N, 9.22 N, and 5.75 N, respectively. The resulting finger postures were also examined using objects of different geometries and sizes.

0.38

Terrain Traversal Without Motion Priors: Expert-Routed Humanoid Locomotion at Deployment Fidelity

cs.RO ·2026-10-07 ·Thanh Ly, Chien Le, Truong-Duy Dang, Duy M. H. Nguyen, Duy-Dung Le +2 more
hot direction
Abstract · https://fai-rl-tech.github.io/moe-perceptive/

State-of-the-art perceptive humanoid locomotion is mostly built on retargeted motion capture, a generative motion prior, and a distillation stage. We report a complete system that does not require this pipeline for a humanoid to cross varied terrain. In its place, we provide a five-expert mixture routed by a terrain classifier over a sensed elevation grid, and target every foothold with a model-based planner, refined by six closed-form-gradient steps on an interpolated terrain-cost field instead of picked from a fixed grid. Driven through the same perception-to-actuation software a robot would run, inside a second, deployment-fidelity simulator, it crosses an $85$\,m five-segment waypoint course $18$ of $20$ times end-to-end ($90\%$, Wilson $95\%$ CI $70$--$97$) and succeeds $20/20$ on four of five segments scored alone, where an ablated configuration without the mixture and without refinement reaches $0\%$ on both the stair and the hurdle segment. The continuous refinement moves $72.4\%$ of foot placements across at least one grid-cell boundary, which produces real sub-grid correction. We also ran matched component studies to report what does not account for it: one expert equals five on the training distribution, four label-free routers collapse to near-uniform activation, and a self-supervised terrain embedding loses to the privileged label.

0.37

Informationally Decoupled Trajectory Design for Sim-to-Real System Identification

cs.RO ·2026-10-07 ·Sangwoo Shin, Ashvin Anilkumar, Ryan Gao, Josiah P. Hanna
code/data
Abstract · arxiv_r2

Sampling-based system identification estimates physically meaningful parameters by tuning a simulator to reproduce the target system dynamics, providing an interpretable approach to improving sim-to-real transfer. Yet when the collected trajectories do not distinguish the effects of different parameters, multiple parameter combinations can reproduce those trajectories, leading to unreliable parameter estimates. To address this challenge, we introduce an Informationally Decoupled Trajectory Design framework (IDTD), which formulates the objective for the exploration policy built on the Schur complement score derived from the Fisher information matrix. To faithfully reflect parameter separability in the exploration objective, IDTD normalizes the score against the per-parameter information, selects the most favorable trajectory segment for each parameter, and aggregates the resulting scores logarithmically. The optimized trajectory is therefore composed of complementary intervals, each exposing a distinct subset of parameters whose contribution to the motion over that interval can be attributed unambiguously. Across diverse simulation environments, ranging from a linear-dynamics system to the Go2 quadruped, G1 humanoid, and Crazyflie quadrotor, IDTD reduces the parameter identification error by 39.6% on average relative to the strongest prior active-exploration baseline and attains improved downstream policy transfer. Furthermore, we validate the proposed trajectory design on a real K1 humanoid, demonstrating that the resulting identified parameters accurately capture the real-system dynamics. Code is available at https://github.com/jsw7460/IDTD.

0.37

Predicted Futures Are Not Enough: Learning Executable Goals for Robot Manipulation

cs.RO ·2026-10-07 ·Tzu-Yu Chuang, Ching-Hsiang Chang, Yi-Hsiu Lee, Yi-Ting Chen, Min Sun +1 more
code/data
Abstract · 9 pages, 8 figures, 4 tables. Project page: https://claire0730.github.io/executable-goals/ Code and models: https://github.com/Claire0730/ex…

Generative world models provide rich predictions of how manipulation scenes may evolve toward task objectives, yet those futures do not directly expose the compact task variables required by control. When training supervises future prediction alone, terminal goal accuracy is not an explicit learning objective, even when geometric recovery is available. We present Entity-Level Goal Readout, a learned prediction-to-execution interface that makes the executable terminal goal an explicit output of a 3D trace world model. It combines object-centric pose prediction with translation grounded in observed depth to produce a compact goal in SE(3). A shared Pose-Native Executor consumes this fixed goal with online object-pose feedback for closed-loop control without rerunning the world model. Across five manipulation tasks, the pipeline achieves a mean success rate of 79.69%. Goal diagnostics directly measure terminal goal accuracy, while controlled translation perturbations characterize how execution degrades under goal error. Zero-shot deployment on a Franka arm achieves 73.33% success on nominal StackCube, 66.67% with distractors, and 75.00% on PickPlate with a target unseen during policy training. These results support treating the prediction-to-execution interface as an explicit learned component of world-model planning rather than incidental post-processing in the control pipeline itself. Project page: https://claire0730.github.io/executable-goals/

0.37

NeRFifyMesh: Optimizing Neural Radiance Fields from Textured Meshes for Robotics Scene Building

cs.RO ·2026-10-07 ·Nillan Nimal, Mahboubeh Asadi, Sajad Saeedi
published
Abstract · Accepted to IROS2026

In robotics, scene representation plays a pivotal role in understanding and interacting with the environment. The advent of Neural Radiance Fields (NeRF) and its variants, as a novel representation, has opened a new frontier of research. In applications such as semantic mapping and simulation, roboticists aim to build scenes using multiple NeRF models, each representing an object. While extensive datasets of 3D mesh models already exist, there is an urgent need to develop tools to convert these assets to NeRF models for rapid algorithm development and testing. This paper presents a new pipeline for converting existing mesh models to NeRF representations by artificially generating a ground truth point-based radiance field through sampling mesh geometry and texture. This approach alleviates the need for camera-based sampling or rendering multi-view images of the original mesh to train the NeRF model. Extensive benchmarking demonstrates that our method yields comparable rendering quality to the baselines. Additionally, the application of this representation is shown by constructing unified NeRF scenes and performing collision simulations with extracted geometry.

0.37

TempoBridge: Language-Guided Tempo Control for Vision-Language-Action Policies

cs.RO ·2026-10-07 ·Yeonseo Lee, Hyosup Shin, Guebin Hwang, Sungho Jo
code/data
Abstract · 8 pages, 5 figures. Project page: https://lysees.github.io/tempobridge-page/

Vision-Language-Action (VLA) models are effective at understanding what task to perform, but provide limited control over how it should be executed, such as moving quickly or slowly. We introduce TempoBridge, a lightweight framework that uses frozen VLA representations to modulate actions according to tempo cues in the instruction at each task phase, without additional tempo-conditioned robot demonstrations or tempo-specific base-policy fine-tuning. TempoBridge extracts tempo cues from contextual VLM representations, aligns them with task progress through a causal phase router, and modulates nominal motion commands during execution. Across LIBERO tasks, TempoBridge improves Tempo Success Rate from 52.6% to 89.7% under canonical tempo instructions while retaining high task success. It also preserves near-baseline performance when no tempo cue is present and generalizes to unseen tempo expressions without additional training. Experiments on a physical robot further demonstrate language-conditioned tempo modulation in real-world manipulation.

0.36

Fast Angular Sweep Monostatic Radar Cross Section Computation via Model Order Reduction

eess.SP physics.comp-ph ·2026-10-07 ·Noelia Naranjo, Valentin de la Rubia
code/data
Abstract

The radar cross section (RCS) is a fundamental parameter in radar engineering, as it determines the detectability of a given target by a radar system. Increasing interest in very low observability (VLO) targets has arisen recently. As a result, especial effort is put in developing numerical tools to accurately predict the electromagnetic scattering from perfect electric conductor objects. Integral equation methods are commonly used in this task. However, full-wave computation of the monostatic RCS at a single angle can be rather time-consuming, nevermind computing the RCS in a specific angular interval with a fine sampling. In fact, this is what is needed in industrial applications. A reduced-order model for fast full-wave monostatic RCS evaluation is proposed to easily achieve fine details in the desired angular domain. An open source integral equation solver, taken into account as a black box, is used to get the electromagnetic scattering in non-penetrable objects at some specific incident angles. Several radar targets will show the possibilities and capabilities of these model order reduction approaches.

0.34

Making Task Abstractions Executable: Control-Aware Layout Repair for a Fixed Controller

cs.RO ·2026-10-07 ·Chiyoung Kim, Seungyeon Back, Sumin Shim, Doyoung Heo, Rita Singh
Abstract · 17 pages, 8 figures, 10 tables

A task abstraction can specify the intended events while its spatial layout prevents a fixed agent and controller from completing them. Starting from a supplied structured task record, we compile whole-task tracking, clearance, and actuation requirements into auditable affine layout constraints. We repair only declared continuous coordinates, preserving event order, timing, topology, and the controller. A most-violated-row update admits conditional finite-certification and net-displacement bounds; a same-compiler quadratic projection separates the representation from the optimizer. On three researcher-authored task abstractions, both backends certify all three layouts and complete all 300 fresh paired rollouts per backend. A risk-target sweep also exposes fixed event tests that the chosen certificate cannot satisfy through layout edits alone.

0.33

Distributed Motion Planning for Multi-Robot Systems under Topological Constraints

cs.RO ·2026-10-07 ·Gianpietro Battocletti, Dimitris Boskos, Dimos V. Dimarogonas, Bart De Schutter
Abstract · 19 pages, 13 figures

Efficient and distributed coordination of mobile robots is one of the main challenges in multi-robot systems. Topological constraints, often expressed as topological braids, are a popular tool to encode complex coordination patterns between multiple mobile robots, as they offer a compact and abstract representation of the desired qualitative relation between the space-time trajectories of the robots. However, execution of joint motion plans encoded as braid-based topological constraints via distributed controllers is challenging, with existing approaches, generally based on the execution of one braid generator at a time, producing slow and suboptimal trajectories. We propose a distributed controller based on Model Predictive Control (MPC) to efficiently execute braid-based topological specifications. Rather than directly tracking the braid specification, we propose to use winding numbers, which are topological invariants for braids, as a proxy. This has the twofold benefit of converting braids into a continuous function, which can be easily tracked by an MPC controller through an appropriate term in the cost function, and of decoupling the global braid specification into a set of pairwise specifications, which can be tracked distributedly through the solution of only local MPC problems. To maintain global coordination, we propose a consensus-based progress estimation approach, which allows the robots to synchronize their motion toward the desired specification. We validate the proposed approach in simulation and in real-world experiments, where we demonstrate the effectiveness of the proposed approach and the improvement over existing approaches in terms of execution speed and control effort.

0.33

LACE-CRAFT: Robot Co-Design with Actor Inheritance and Blackboard Collaboration

cs.RO ·2026-10-07 ·Yuhan Wen, Jiawei Wang, Qixuan Zhang, Calvin Zhang, Yusen Qin +1 more
Abstract · 16 pages including appendix. Corrected source files to match the author list; scientific content unchanged. Project website: https://deemost…

Robot co-design couples morphology search with policy learning, yet training every new design from scratch discards acquired control experience. We present LACE-CRAFT, which compares continued learning on the current robot with policy adaptation to new morphology-reward pairs. LACE resumes the incumbent's full learning state and initializes compatible challengers with its actor parameters and observation statistics. A fixed task metric selects among both branches and the frozen incumbent. CRAFT coordinates Feedback, Morphology, Reward, and Integration roles through shared experimental records and behavioral replays to generate and cross-review paired proposals. A generative extension converts generated meshes into editable articulated models with configured joints, actuator interfaces, and consistently updated simulation assets. Across five locomotion benchmarks, mean scores over three evaluation seeds are 6.4-91.9% higher than D2C. Both methods train 30 new morphology-reward pairs over five rounds; LACE additionally uses four continuation training units. Five-task ablations examine policy inheritance and replay-derived feedback. A fabricated prototype demonstrates indoor walking and illustrates the geometry-to-hardware workflow.

0.33

Deceptive Stochastic Patrolling via Markov Chain Lifting

eess.SY math.OC ·2026-10-07 ·Yohan John, Gilberto Diaz-Garcia, Jason R. Marden, Francesco Bullo
theory
Abstract · 14 pages, 8 figures. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after w…

In this paper we propose lifted Markov Chains (MCs) as a new paradigm for deriving patrol strategies for mobile agents on an environment represented as a graph. Lifted MCs operate on a state space that can be larger than the set of nodes of the graph, and a projection operation maps the MC state to the corresponding node of the graph. We prove bounds on the performance improvement of lifted MCs over non-lifted MCs, where the performance is measured by the Kemeny constant, and show a corresponding monotonicity for the Stackelberg game capture probability and return-time entropy. These bounds yield a result of independent interest that relates the Kemeny constant and the MC conductance for general MCs. We also provide a tractable method for optimizing lifted MCs on general graphs with edge weights encoding travel times. Simulation results on randomly generated and real-world graphs verify that lifted MCs achieve substantial improvements, particularly on sparse graphs.

0.33

Characterizing the Narrowband Ambiguity Function of Frequency Modulated OFDM Waveforms

eess.SP ·2026-10-07 ·David A. Hague
published
Abstract · Accepted for presentation at the IEEE ISAC Conference

This paper characterizes the spectral and Ambiguity Function (AF) properties of Frequency Modulated Orthogonal Frequency Division Multiplexing (FM-OFDM) waveforms employing Phase-Shift Keying (PSK). The mainlobe structure of the AF is well approximated by the Ellipse of Ambiguity (EOA) model. The EOA produces exact closed-form expressions that describe the AF mainlobe width in time-delay and Doppler as well as the degree of coupling between them. These expressions yield three fundamental insights into the spectral and AF shapes of FM-OFDM waveforms employing PSK: (1) these waveforms possess a fixed Root Mean Square (RMS) bandwidth for a given modulation index $h$, pulse-length $T$, and number of sub-carriers $L$, (2) they are roughly twice as spectrally efficient as Constant Envelope OFDM (CE-OFDM) waveforms with the same RMS bandwidth, and (3) for practical waveform design parameters will always possess a ``Thumbtack-Like'' AF shape. Finally, this paper demonstrates a waveform optimization technique that minimizes the FM-OFDM waveform's Auto-Correlation Function (ACF) sidelobes using only a subset of the sub-carrier symbols as sensing ``pilot'' symbols leaving the rest for communications. This pilot optimization technique is then demonstrated via two illustrative design examples.

0.32

Targeted Modality Dropout for Real-Robot Manipulation Robust to Intermittent Vision Loss

cs.RO ·2026-10-07 ·Genki Shikada, Kazuki Osamura, Masaru Ide, Tetsuya Ogata, Kanata Suzuki
Abstract · 7 pages, 5 figures, 2 tables. Submitted to the 2027 IEEE/SICE International Symposium on System Integration (SII 2027)

Imitation learning policies that integrate multiple sensory modalities are prone to overreliance on a dominant modality, such as vision, during training, which can disrupt policy execution when that modality is lost at inference time. In this paper, we introduce Targeted Modality Dropout (TMD), in which the dependence on each modality is estimated using attention and the most dominant modality is selectively dropped. This is combined with entropy regularization over the dependence distribution. Through real-robot evaluation using a bimanual manipulator, we show that under vision loss the success rate of the baseline policy drops substantially, whereas TMD sustains task execution. In contrast, a conventional dropout that selects the dropped modality at random, without the entropy regularization, fails on many tasks even without vision loss.

0.32

RealtimeWAM: How Fast Can I Run My World Action Model?

cs.RO ·2026-10-07 ·Huanan Liu, Ye Li, Kangye Ji, Xiaoyu Chen, Hanyun Cui +6 more
code/data
Abstract · 24 pages, including appendix. Project page: https://anonymous.4open.science/w/realtimewam/

World Action Models (WAMs) combine visual dynamics modeling with action generation, but their high inference latency limits responsive robot control. Recent efforts accelerate inference by removing explicit future-video generation at test time, as in FastWAM, an approach that requires a specially tailored architectural design. More general caching strategies exploit feature redundancy, but redundancy alone does not capture the changing computational demands of closed-loop control. To address these challenges, we present RealtimeWAM, a general, training-free framework that coordinates parallel execution with adaptive computation for low-latency inference across diverse WAM architectures. We exploit layerwise dependencies to overlap observation processing with prediction. However, concurrent branches still compete for GPU resources, limiting the benefit of parallel execution. We therefore adapt computation throughout the pipeline through selective reuse, caching observation features in visually stable regions and reusing Transformer residuals while reserving additional refinement for small predicted adjustments. We evaluate RealtimeWAM on FastWAM and OpenWAM across RoboTwin, LIBERO, and LIBERO-Plus. On an RTX 4090, measured mean inference latencies are 24.09 and 63.09 ms, corresponding to average speedups of 8.90$\times$ and 10.67$\times$. Average success rates are 82.75% and 87.41%, respectively, within 0.02 and 0.53 percentage points of native inference. Across five real-world tasks, RealtimeWAM improves average success rates over native inference by 17.2 and 37.2 percentage points on FastWAM and OpenWAM, respectively.

0.32

Finite-Aperture Modeling of Pattern-Reconfigurable Antennas for Multiuser Beamforming

eess.SP ·2026-10-07 ·Kangjian Chen, Han Qing Yang, Hang Wong
Abstract · 14 pages, 11 figures; submitted to IEEE for possible publication

This paper establishes a finite-aperture model for pattern-reconfigurable antennas (PRAs) and exploits its aperture-constrained Fourier representation to develop an efficient multiuser beamforming framework. First, we develop an ideal model by representing the complex radiation pattern using a Fourier series. By applying the Jacobi-Anger expansion, we relate the antenna aperture to the effective Fourier order and derive aperture-dependent power bounds for individual Fourier coefficients. Second, we employ a practical multiport network model and design three representative reconfigurable pixel antennas to verify the predicted effective Fourier orders and per-order power constraints. We then formulate weighted sum-rate maximization for multiuser downlink transmission under both models. Using bounded Fourier coefficients as auxiliary variables, we develop a coefficient-bound penalty-based weighted minimum mean-square error algorithm that separates continuous beamforming from discrete switch-state optimization. We further propose a one-bit neighborhood search that exploits the Sherman-Morrison identity to accelerate the discrete optimization. Simulation results validate the physical relevance of the ideal model and demonstrate multiuser sum-rate gains and reduced optimization runtime for densely switched antennas. The proposed model provides a physically grounded basis for PRA-enabled communication analysis and design, while the framework accommodates different beamforming objectives and PRA implementations.

0.32

Point It, Strike It: Direction-Conditioned Dynamic Manipulation of Deformable Linear Objects

cs.RO ·2026-10-07 ·Yi Yang, Xiang Fei, Lehong Wang, Zilin Dai, Ruogu Li +7 more
code/data
Abstract · 9 pages, 5 figures, 4 tables. Project Page: https://deformx.github.io/DeformY

Goal-conditioned dynamic manipulation of deformable linear objects has mainly specified goals as positions for a rope tip to reach. Many tasks, however, depend on how the tip arrives. We therefore study single-swing rope striking with goals that specify the tip's 3D position and arrival direction, across the workspace and on different ropes. This is challenging because rope dynamics are hard to model, no demonstrations exist, distinct swings reach the same goal with different reliability, and the sim-to-real gap extends beyond the rope. To address these challenges, we extend the state-of-the-art DLO simulator DeformX with GPU acceleration, a stable Cosserat rod solver, and a cross-flow aerodynamic model, yielding DeformX2.0, which is more than $20{,}000\times$ faster. We then propose TRACE (Trace-rooted Adaptive Cross-Entropy), which generates striking data by warm-starting each new target from the stored swing whose tip path passes closest to it. Its cost penalizes rope bending and abrupt tip motion to favor repeatable swings. A conditional flow-matching policy trained on this data reaches 92.1% accuracy in simulation. Finally, we propose RECAP (Residual Calibration Policy), which fits the simulator's rope and rig parameters to a few calibration swings and adapts actions with a correction policy trained in simulation. On a real robot, across three ropes, RECAP raises success within 5cm from 72% to 87% for position goals, and within 10cm and 10° from 50% to 79% for goals that also specify the arrival direction.

0.31

Video-to-Model: Automatic Modeling of Deformable Linear Objects

cs.RO eess.SY ·2026-10-07 ·Akshun Sharma, Kimia Forghani, Yancy Diaz-Mercado
Abstract · 14 pages, 5 figures

This paper presents a video-to-model framework for automatically modeling the motion of a deformable suture thread from an input video. We utilize a recently developed CBF--CLF--QP numerical model that simplifies the characterization of deformable string motion through the selection of a small number of parameters. A perception module first localizes and tracks the thread in video, producing an ordered sequence of thread nodes. The observed thread motion is then processed by a spatio-temporal CNN network that estimates the effective parameters of a structured CBF--CLF--QP model. These parameters are used to simulate the thread under a user-defined needle velocity input. Experiments using unseen thread configurations and motion demonstrate that the framework can reliably reconstruct the thread behavior from video, automatically configure the structured model, and reproduce the expected thread motion with low tracking error. The proposed approach reduces the need for manual parameter tuning and provides a step toward automatic video-based modeling of deformable linear objects.

0.31

Antenna Coding by Pixel Antennas: Unlocking A New Dimension for Wireless Communications

eess.SP cs.IT ·2026-10-07 ·Zixiang Han, Shanpu Shen, Yifei Yuan, Jiangzhou Wang, Ross Murch +1 more
Abstract

Conventional wireless communication systems employ antennas with fixed radiation patterns. Enhancing system performance by increasing the number of antennas would push the limits of cost and size for communication devices. To overcome this issue, novel antenna technologies are urgently desired to break through the fundamental performance bottleneck and engineering limits. Pixel antennas provide a general framework for highly reconfigurable antennas, which can flexibly adjust radiation patterns via controlling variable lumped elements placed across adjacent pixels, unlocking novel degrees of freedom (DoF) in the beamspace domain. Recently, an emerging antenna coding technique links wireless system performance to the antenna configurations. This enables the use of pixel antennas to enhance system performance by leveraging the DoF from the antenna side. In this article, we provide a comprehensive overview on antenna coding by pixel antennas. Specifically, based on the pixel antenna model, the concept of antenna coding is firstly introduced, followed by a discussion on the classifications of antenna coding. Two cases for channel gain and multiplexing gain enhancement are highlighted to show the effectiveness of antenna coding in wireless communications. Besides, various optimization approaches for antenna coding design are investigated. Finally, the challenges and opportunities are outlined, providing guidance for future works.

0.31

Fast Antenna Coding Based on Pixel Antennas for Enhancing Spectral Efficiency of MIMO Systems

eess.SP ·2026-10-07 ·Zixiang Han, Xin Su, Jing Jin, Shanpu Shen, Qingqing Wu +3 more
Abstract

A fast antenna coding technique based on pixel antennas is investigated. Antenna coding is a technique that dynamically controls the ON/OFF states of PIN diodes to adjust radiation patterns of pixel antennas. By switching the antenna coder at the baseband symbol period, this technique can modulate information on the radiation patterns in the beamspace. This allows the pixel antenna to transmit additional information, thereby enhancing the spectral efficiency (SE) of multiple-input multiple-output (MIMO) systems. Specifically in this work, we firstly introduce the concept of pixel antenna and the fast antenna coding. Then based on the beamspace channel representation, we provide the system model of MIMO with the fast antenna coding and derive the SE and bit error rate (BER). The SE enhancement achieved by fast antenna coding is also demonstrated. An optimization approach is also presented for obtaining the antenna coders that can generate the desired radiation patterns for fast antenna coding. Using the optimized antenna coders, we evaluate the SE, energy efficiency (EE), and BER of MIMO systems with fast antenna coding. The simulation results show that for a 4-by-4 MIMO system, SE is enhanced by 12 bits/s/Hz and EE can be improved by 33\% when compared with conventional MIMO, demonstrating the effectiveness of the fast antenna coding technique.

0.31

Co${}^{2}$Skill: Whole-Body Control via Skill Composition for Long-Horizon Human-Environment Interaction

cs.RO cs.GR ·2026-10-07 ·Jeonghwan Kim, Hyeonwoo Kim, Hanbyul Joo
code/data
Abstract · 12 pages, 3 figures, 2 tables. Project page: https://canoneod.github.io/projects/coskill/

Achieving human-level dexterity in complex, unstructured environments requires the seamless integration of whole-body scene interaction and dexterous object manipulation skills. While existing physics-based controllers generate physically plausible behaviors in each domain, they largely address these two capabilities independently. In this paper, we present Co${}^{2}$Skill that integrates scene interaction and dexterous manipulation through a unified policy formulation. Built on a pretrained motion prior, the policy uses task and phase dependent observation masks to select information relevant to the current interaction goals. We introduce a goal-conditioned loco-manipulation curriculum that combines partial reference guidance for precision with exploration from varied initial states while allowing goal-directed execution beyond the demonstrated trajectories. We further introduce a cross-task curriculum that jointly trains individual skills and selected task sequences, preserving physical states across task boundaries and maintaining grasps during subsequent scene interactions. Together, these support sequential task execution and simultaneous scene interaction with object manipulation. We evaluate sitting, standing, climbing, stair traversal, and goal-directed manipulation, together with sequential execution and with random different conditions. Additionally, we demonstrate skill compositions in indoor environments, illustrating their integration within the same control formulation.

0.31

Same Action, Different Outcome: Variability in Dynamic Cloth Manipulation

cs.RO ·2026-10-07 ·Mahed Dadgostar, Guillem Alenyà, Júlia Borràs
Abstract

Although cloth is known to exhibit different outcomes under repeated fast dynamic motions, even when the same trajectory is applied, this variability has not yet been systematically characterized. Quantifying it is essential to assess the reliability of learned manipulation policies and the extent to which simulation can reproduce real-world behavior. To study this, we execute the same trajectory ten times across four dynamic tasks, two of which are novel, each tested with three cloths of very different properties and at up to three execution speeds, with a total of 269 recorded rollouts. For all of them, we record small marker positions on the cloth and synchronized stereo camera. We then formalize different metrics to quantify variability, and our results show how it is significant in every test condition, in most cases one to three orders of magnitude above the repeatability inherent to the robot and sensing noise. Our results also show variability is driven mainly by the cloth physical properties but also grows with speed. By replaying all the rollouts in four calibrated modern cloth simulators, we show that none of them can reproduce the variability magnitude we observed in real cloth, nor its ordering. Together with our analysis, we publish the dataset with synchronized OptiTrack, vision and robot logs and their corresponding simulator twins.

0.31

A Single-Inductor N:1 Resonant Switched-Capacitor Ladder Converter with Efficient Continuous Voltage Conversion Capability

eess.SY ·2026-10-07 ·Sandeep Reddy Kukunuru, M Mahmudul Hasan Sajeeb, Rohit Reddy Karnaty, Loai G. Salem
Abstract

This paper presents an N:1 resonant switched-capacitor (ReSC) ladder converter of an output voltage that can be regulated continuously below the native N:1 resonant ratio while maintaining high efficiency. The topology can be generalized to any N-to-1 dc-dc conversion application while requiring only a single inductor. Efficient wide-range voltage regulation is realized by operating the bottom stack of the ReSC switches as diodes to maintain their switching always in phase with the inductor current, and hence, circumvent circulating currents that increase the inductor RMS current and ohmic losses under frequency control. A theoretical model has been developed to predict the output voltage and efficiency of the converter, with simulation results confirming its accuracy for N = 1, 2, and 4. A hardware prototype on a printed circuit board is developed to validate the performance of the proposed frequency-controlled regulation scheme. The prototype achieves a peak efficiency of 97% (94.3%) with a load of 30Ohms (resp. 6Ohms) and an efficiency >90% while generating a 9V USB-compatible voltage from a 40V-to-60V input. The measurement results match the model predictions within 0.63% accuracy.

0.30

Centrality-based Structural-Electrical Contingency Screening and Transmission Reinforcement Prioritization

eess.SY ·2026-10-07 ·Abanish Tiwari, Chandan Chaudhary, Mekh Raj Joshi, Mohammed Ben-Idris, Joydeep Mitra
published
Abstract · Accepted for IAS Annual Meeting 2026, Vancouver, Canada

Rapid contingency screening is essential for secure transmission system operation and planning. Although full AC contingency analysis provides accurate system assessment, it requires high computational effort for large contingency sets. Topology-based screening methods reduce this burden but may not accurately represent operating-point-dependent electrical stress. In addition, conventional single-contingency screening does not quantify interactions among simultaneous outages or provide a direct criterion for reinforcement prioritization. This paper presents a hybrid structural-electrical screening approach that combines a computationally efficient, dispatch-independent structural analysis based on reactance-weighted centrality measures with a flow-based severity assessment using line outage distribution factors. Two indices are evaluated using the flow-based severity: a synergy index to quantify super-additive interactions among simultaneous line outages and a tolerance index to rank reinforcement candidates based on their reduction in aggregate post-contingency severity. The correspondence between structural indicators and electrical severity is evaluated quantitatively. Tests on the IEEE 39-bus system identify 115 super-additive line pairs beyond a 5% threshold, with the most severe interaction exceeding seven times the linear combination of individual outage effects. The highest-ranked reinforcement reduces the mean post-contingency thermal severity by 7.5%. The structural rankings show limited correlation with flow-based severity, with a Spearman coefficient of approximately 0.2. The proposed approach provides an efficient pre-screening method to reduce contingency and reinforcement search spaces before detailed AC and economic evaluations.

0.30

SiGNgapore - An Interactive Dataset for Sign-based Visual Navigation

cs.RO ·2026-10-07 ·Nicky Zimmerman, Joel Loo, Zishuo Wang, David Hsu
benchmarkdataset
Abstract · under review for IJRR

Robots navigating open-world, human-centered environments must look beyond prebuilt high-fidelity maps and leverage navigational aids embedded in their surroundings, particularly, signs. To advance sign-guided visual navigation, we introduce SiGNgapore, an interactive, multimodal dataset designed to benchmark sign-guided multi-step navigational decision-making. SiGNgapore covers diverse public spaces in Singapore, including shopping malls, public transportation hubs, and hospitals, featuring multi-floor, multi-building layouts and indoor-outdoor transitions. It consists of 56 long-horizon navigation missions and 456 scenes, each containing a navigational sign and its surrounding environment. We provide synchronized RGB, depth, IMU, and odometry data for each scene. We additionally provide the layout map and the GPS-aligned scene graph for selected mission sites. Beyond navigational decision-making, SiGNgapore supports a variety of spatial-semantic reasoning tasks, including navigational sign parsing and visual sign grounding, with the ultimate aim of driving progress in mapless robot navigation.

0.30

IVG-UAV: An Intelligent Voice-Guided UAV System for Autonomous Ripe Fruit Harvesting with Vision-Based Classification and Adaptive Path Planning

cs.RO ·2026-10-07 ·Dinh Trung Duong
Abstract

In tropical regions, their agricultural sectors remain highly dependent on manual labor for fruit harvesting. On large-scale farms, this dependency often results in significant labor cost and logistic complexities. This project presents the development and simulation of a voice-controlled Unmanned Aerial Vehicle (UAV) system designed to automate harvesting tasks in extensive plantations. The proposed system integrates speech recognition using Whisper [1] and LLM, computer vision-based ripeness classification, and adaptive path planning within a unified framework. The entire system is modeled and validated in a Gazebo simulation environment, allowing performance evaluation under controlled agricultural scenarios

0.30

On-Demand Robotic Assembly via Differentiable Geometric Part Repair

cs.RO ·2026-10-07 ·Millicent Schlafly, Fabio Schaub, Diogo Costa Pais, Luca Lelli, Janne Dvorak +3 more
Abstract · 8 pages, 4 figures, and 2 tables

Transitioning from a digital design to a robotic assembly process currently requires months of expert manual tuning to reconcile part geometries with robotic constraints. This paper presents an end-to-end, autonomous pipeline for the design and physical construction of bespoke wooden assemblies. A generative AI agent translates user prompts into initial 3D geometries, balancing the visual fidelity of the design with select physical constraints. The assemblability of the design is further improved by a gradient-based repair stage that backpropagates through a graph attention network surrogate to adjust component geometries. In addition to correcting for disjointed and overlapping components, we demonstrate hardware-specific corrections, differentiably optimizing the geometry of components to enable robot screwdriving for 86.7% of 60 novel natural language inputs, significantly outperforming prior work by a factor of ten. For ten of the structures, we physically demonstrate assemblability with two UR5e robots. This work marks a meaningful step toward on-demand robotic manufacturing, enabling the rapid production of customized, low-volume goods.

0.30

Low-complexity Equalization of Zak-OTFS Via Neumann Series

eess.SP cs.IT ·2026-10-07 ·Vineetha Yogesh, Saif Khan Mohammed, Sandesh Rao Mattu, Ronny Hadani, Robert Calderbank
Abstract

We describe a general method for selecting an orthonormal basis of carrier waveforms that aligns the basis with delay / Doppler characteristics of a wireless channel. We show that our method enables low-complexity equalization for two channels of practical interest. The first is satellite communication and the second is communication from a ground station to an unmanned aerial vehicle (UAV). After Doppler compensation both scenarios are characterized by a first line of sight (LOS) path with zero delay and zero Doppler shift, and a second weaker path. We show that our method is robust to fractional delay and Doppler shifts. We consider orthonormal bases of carrier waveforms that are obtained from the pulsone basis of Zak-OTFS carrier waveforms by applying a generalization of the discrete affine Fourier transform (GDAFT). This family of waveforms includes AFDM and other modulations proposed for 6G. What distinguishes these bases is that the carrier waveforms are common eigenvectors of some maximal commutative subgroup S of a Heisenberg-Weyl group of discrete delay and Doppler shifts. We describe how to choose S to mitigate the damaging effects of interference between carriers. We represent the wireless channel by delay-Doppler taps and observe that a channel tap located within S multiplies every waveform by a complex phase. If all channel taps are located within S, then the channel multiplies every waveform by a complex phase, and a single tap equalizer supports reliable communication. This is the case for a linear time-invariant (LTI) channel, where S is the group of discrete time shifts and the carrier waveforms are discrete tones (OFDM). In general, we choose the subgroup S to maximize the diagonal component of the channel energy.

0.30

FoldBack: Self-Correcting Masked Generative Policy for Long-Horizon Garment Folding

cs.RO cs.AI ·2026-10-07 ·Lipeng Zhuang, Shiyu Fan, Yingdong Ru, Zhuo He, Florent P. Audonnet +2 more
hot direction
Abstract

We present FoldBack, a self-correcting masked generative policy for long-horizon garment folding. Existing long-trajectory policies may continue after a missed or slipped grasp even when the garment has not reached the intended configuration. We structure FoldBack's recovery mechanisms around three inference-time decisions: when to refine and verify, how to roll back, and where and how to retry. FoldBack aligns refinement and grasp verification with pick-and-place events, returns the robot to a retryable pre-grasp configuration while preserving successful grasps, and selectively regenerates the failed segment and selected future actions while avoiding previous failed grasp locations. To our knowledge, FoldBack is the first editable full-trajectory policy to unify these decisions, enabling failed interactions to be detected, undone, and repaired before execution continues, without recovery demonstrations or base-policy retraining. Across 33 real garments from six categories, FoldBack achieves 75.2% final folding success and 0.837 final-mask IoU, versus 45.7% and 0.689 for the strongest prior baseline.

0.30

Adaptive Code Generation for Controlling Robots

cs.RO cs.AI ·2026-10-07 ·Justus Flerlage, Thorsten Wittkopp, Alexander Acker, Odej Kao
published
Abstract · International Symposium on Leveraging Applications of Formal Methods, Verification, and Validation 2026

Deploying robots as Complex Adaptive Systems (CAS) in unknown and dynamic environments necessitates a transition from rigid command libraries toward intention-based autonomy, as natural language represents the only medium capable of articulating complex goals beyond the capacity of finite instruction sets. While Large Language Models (LLMs) offer a path toward natural language goal description, their integration introduces significant challenges: the formalization gap between imprecise intentions and executable actions, the taxonomy gap induced by unpredictable environments, and the challenge of maintaining temporal state and progress awareness. This work introduces an architectural framework that enables robotic control by leveraging generative AI. The system follows a dual-AI design: an LLM translates high-level intentions into executable program code restricted to a formal robotic library and constrained by verifiable syntax, while a Vision-Language Model (VLM) provides semantic grounding via a distillation process. To ensure robustness, the framework incorporates environment-driven replanning triggers based on geometric and semantic thresholds, complemented by continuous runtime monitoring and an adaptive planning loop. Benchmarked across frontier models, our framework architecture demonstrates that grounding generative AI in a reactive, constrained loop enables robust fulfillment of complex intentions in dynamic and unknown environments.

0.30

Precise SE(3) End-Effector Tracking in Whole-Body Humanoid Control

cs.RO eess.SY ·2026-10-07 ·Joohwan Seo, Xiaofeng Guo, Jinkun Cao, Roberto Horowitz, Rocky Duan +2 more
Abstract

Precise end-effector tracking during humanoid whole-body motion is challenging due to floating-base oscillations, gravity, dynamic coupling, and locomotion-induced disturbances. We propose ResGAC, a whole-body humanoid controller for precise end-effector pose tracking that combines geometric admittance control (GAC) with residual reinforcement learning. GAC provides structured $\SE$ task-space feedback and generates nominal arm joint-position targets, while residual RL compensates for unmodeled dynamics and coordinates locomotion and balance in the shared joint-position action space. The left-invariant geometric formulation allows the same GAC law to be used across manipulation reference frames. This enables the use of a ground-attached heading frame that preserves planar locomotion while removing pelvis roll, pitch, and heave from the manipulation reference, thereby reducing reference-induced end-effector motion during locomotion. ResGAC is validated on a real Unitree G1 humanoid. Across four standing end-effector tracking benchmarks, ResGAC consistently outperforms representative baselines, including SONIC, achieving lower translational and rotational errors. Real-world experiments further demonstrate reduced propagation of pelvis motion to the desired end-effector pose using the proposed ground-attached heading frame. ResGAC achieves $90\%$ success in a standing peg-in-hole task compared with $50\%$ for SONIC, and accurate world-frame $\SE$ end-effector pose tracking during lower-body motion. Experimental videos are included in the supplementary material and are also available on the project website: https://resgac.github.io/ResGAC-website/.

0.29

LLA-MPPI: Rapidly Adaptive Whole-body Control of Legged Robots with GPU-Accelerated Parallel Simulations

cs.RO eess.SY ·2026-10-07 ·Sebin Jung, Maitham F. AL-Sunni, Juan Alvarez-Padilla, Zachary Manchester, Changliu Liu +1 more
code/data
Abstract · The first two authors are co-first authors (contributed equally to the work)

Real-time whole-body controllers for legged robots typically plan through a fixed nominal model and degrade when the deployed dynamics change. Adaptive methods typically require a model structure that contact dynamics do not provide, or they need offline training for each anticipated condition. We present Look-back and Look-ahead Adaptive Model Predictive Path Integral control (LLA-MPPI). The method converts whole-body adaptation into selection over a bank of GPU-batched contact simulators with different physical or structural parameters. Windowed prediction errors select the simulator that best explains recent motion. A whole-body MPPI planner optimizes controls through the selected model. The framework requires no offline training, and its selected hypotheses are physically interpretable. Across four simulated tasks, it achieves 97.5% success while the strongest baseline reaches 74% and an oracle with the true model reaches 98.5%. Hardware validation on a Unitree Go2 shows the robot walking under a payload added mid-run, walking after one leg is disabled, and pushing a box to its goal while increasing its mass on the fly. Code, videos, and project details are available at: https://lla-control.github.io

0.29

COOL: Curiosity-Driven Object Ownership Learning for Personalized Robotic Assistance

cs.RO ·2026-10-07 ·Samira Huber, Ruben Hammele, Sören Pirk
published
Abstract · Accepted to CoRL 2026

Robots are increasingly expected to provide personalized services in everyday environments. To do so, they must ground natural-language commands such as "Where is my backpack?" or "Find my bottle" and execute them by reasoning about object instances, people, locations, and ownership. This is challenging because ownership is rarely labeled explicitly and must be inferred from long-term, behavioral evidence of human-object interactions. To address this, we present COOL, a novel robotic framework for autonomously learning object ownership from everyday observations and maintaining a long-term spatial memory of its environment. To keep its memory current, COOL uses an agent-based curiosity-driven data collection strategy that guides the robot toward the most promising locations to gain information and refresh stale observations. Offline experiments, ablation studies, and real-world evaluations show that COOL can infer ownership relations from real-world interactions and use this knowledge for ownership-conditioned navigation and task execution.

0.29

Cooperative Dueling DQN SAC Learning for Energy Efficiency in Dynamic OWC Networks

eess.SY ·2026-10-07 ·Walter Zibusiso Ncube, Ahmad Adnan Qidan, Taisir El-Gorashi, Jaafar M. H. Elmirghani
hot direction
Abstract

Growing wireless traffic is increasing pressure on the congested radio-frequency spectrum. Optical wireless communication (OWC) provides a complementary solution by using the abundant unlicensed optical spectrum. However, indoor OWC networks are dynamic: users move, enter or leave the network, and subscribe to different services. Poorly coordinated resource allocation can consequently waste subcarriers, require excessive transmission power and cause frequent AP reassignments. Energy efficiency (EE), defined as the total delivered data rate divided by the total network power consumption, therefore requires the serving AP, number of allocated subcarriers, and transmission power to be jointly adapted while maintaining QoS. Optimising these in a dynamic time series, multi-service OWC environment produces a complex sequential EE optimisation problem. To address this problem, this work proposes Dual-Agent Resource Allocation using Deep Reinforcement Learning (DARA-DRL). DARA-DRL combines a branching duelling deep Q-network for association and subcarrier allocation with a conditional soft actor-critic agent for continuous power control. The agents are coupled through a common reward and a cooperative value update that evaluates each discrete allocation together with its corresponding power decision. Simulation results show that DARA-DRL remains within 5\% of the optimal solution and, compared with state-of-the-art benchmarks, it improves EE by 28.8\% and QoS satisfaction by 10.5\%, while reducing online decision time by 14.5\%. Results demonstrate that agent specialisation simplifies mixed-action learning, and cooperation outperforms independently trained agents.

0.29

Benchmarking Behavioral Steerability in Behavior Foundation Models

cs.RO ·2026-10-07 ·Minghe Gao, Zhanxi Yan, Jiahui Liu, Wendong Bu, Xiaoting Chen +7 more
benchmark
Abstract

Behavior Foundation Models (BFMs) are emerging as a paradigm for translating human intentions into executable humanoid behaviors. As these models evolve beyond behavior generation toward general-purpose behavioral systems, a fundamental question arises: can they be reliably steered according to user intentions? In this paper, we introduce the concept of behavioral steerability, defined as the ability of BFMs to faithfully generate behaviors that satisfy user-specified intentions. To study this capability, we present RoboSteer, the first benchmark for behavioral steerability in BFMs. RoboSteer organizes behavioral steerability into a three-level hierarchy-Conditional Steering, Constraint Steering, and Compositional Steering-and establishes a unified evaluation framework supported by a large-scale multimodal motion corpus. Using RoboSteer, we conduct the first large-scale empirical study of behavioral steerability across 9 existing BFMs. We view behavioral steerability as more than a capability for controlling motion: it concerns how embodied systems translate human intentions into purposeful actions. We hope RoboSteer will advance research on intention realization as a foundation for general-purpose embodied intelligence.

0.29

Diagnosing and Recovering from Observation-Space Shift at Long-Horizon Skill Seams

cs.RO cs.LG ·2026-10-07 ·Pranav Wagh, Yu Fang, Yue Yang, Mingyu Ding
benchmark
Abstract · 8 pages, 4 figures, 7 tables. Submitted to ICRA 2027

Long-horizon robotic manipulation is often built by chaining independently trained skills. Although each skill can be reliable in isolation, performance degrades sharply when skills are chained: each downstream skill must start from the state its predecessor leaves behind rather than from its training distribution. We study this failure mode, Observation-Space Shift (OSS), and ask what causes these skill-seam failures. Using privileged simulator resets, we find that the dominant shift comes from displaced scene state (e.g., an open drawer or secondary objects left behind by earlier skills), not from the robot's joint configuration or the object the downstream skill manipulates. To test this diagnosis, we build a fully learned detect-restore-resume system: a task-progress monitor detects the stall, a learned policy restores the displaced scene components, and seam-robust fine-tuning lets the skill resume. It recovers the seam where every tested alternative fails, which we treat as evidence for the diagnosis rather than as a general-purpose method. On the BOSS-44 benchmark, the system improves full-chain success from 7.6% to 26.5%, a 3.5x improvement over the base policy and 51% of a privileged restoration oracle, whereas best-of-K resampling, a Diffusion Policy, and world-model baselines fail to recover from the evaluated seam states. On a real Franka arm running a fine-tuned $π_{0.5}$ policy, the same monitor is limited by exterior-camera observability, yet closing the loop still recovers some otherwise-terminal failures, motivating wrist and gripper sensing. These results suggest that some long-horizon composition failures are better addressed by restoring the scene before resuming the policy than by retrying from an off-support state.

0.29

Contact-Aware Imitation Learning Through Contact Factorization

cs.RO ·2026-10-07 ·Jiho Hong, Daeun Song, Sanghyun Kim, Mingyo Seo
code/data
Abstract

Generalizable contact-rich manipulation requires robots to preserve intended task behavior while adapting its physical realization to changing contact conditions. However, interaction forces can vary substantially with small changes in surface geometry, orientation, and friction, making policies trained directly on raw force measurements difficult to transfer beyond demonstrated conditions. We introduce FACE, a contact-factorized imitation learning framework that separates intended task behavior from environment-dependent contact factors. Our representation expresses interaction forces in normalized, contact-relative coordinates, while a learned contact-normal estimator and an online friction estimator infer the local contact normal and effective friction scale. Together, these estimators enable force observations to be encoded and policy outputs to be decoded into physical motion and force commands during execution. In this way, FACE adapts execution to current contact conditions while preserving the intended task behavior, without updating the policy parameters. We evaluate FACE on real-robot contact-rich manipulation under unseen variations in surface properties and geometry, demonstrating robust generalization across contact conditions through controlled comparisons with variants that adapt prior approaches to our setting. Videos and additional materials can be found on the project page: https://rcilab.khu.ac.kr/face.

0.28

TAPNAV: Humanoid Navigation through Tactile Active Perception

cs.RO ·2026-10-07 ·Huaze Liu, Zhenyu Wu, Jaehwi Jang, Junjie Sheng, Andrew Collins +6 more
Abstract · 9 pages, 10 figures

Navigation in vision-denied environments is challenging for humanoid robots because proprioceptive odometry drifts and localization uncertainty accumulates rapidly. We present TAPNAV, a tactile active-perception framework that enables humanoid navigation toward a goal by actively probing surrounding structures without relying on vision. TAPNAV maintains a pose belief from odometry, IMU, and tactile contact observations, and couples uncertainty-aware global route planning with information-gain-driven local probing. The global planner searches for routes that keep predicted localization uncertainty bounded by exploiting opportunities for tactile correction, while the local planner selects probe actions that maximize expected information gain. A whole-body controller coordinates the humanoid's locomotion and end-effector contact to execute the planned navigation and probe motions. We evaluate TAPNAV in simulation and on a Unitree G1 across different floor plans and obstacle geometries. TAPNAV achieves lower state estimation error and a higher task completion rate than baselines. These results demonstrate that actively planning physical interactions with the environment can provide localization cues for reliable humanoid navigation without vision.

0.28

RoboPrompt: Intuitive Robot Policy Steering with Sparse Human Input

cs.RO ·2026-10-07 ·Yanwen Zou, Chenyang Shi, Guoxuan Xu, Wenye Yu, Wendi Chen +3 more
Abstract · 8 pages

End-to-end robot policies trained through imitation learning remain constrained by limited data diversity, making reliable zero-shot deployment in real-world settings challenging. Shared-autonomy methods enable human correction through teleoperation, but specialized hardware and operator training hinder deployment at scale. Other approaches incorporate human guidance as additional policy inputs, often requiring architectural changes and dedicated training for steerability, which limits their applicability across policies. We present RoboPrompt, a general-purpose, lightweight robot policy steering system that enables users to guide policy behavior through intuitive, sparse inputs, including drawn traces, target points, and coarse directional instructions. RoboPrompt decouples human-intention translation from the underlying policy: a reusable module converts human guidance into action drafts, which are refined through the diffusion or flow-matching dynamics of the base policy. By controlling action generation in noise space, RoboPrompt balances human intent with the policy prior without modifying the base policy architecture or fine-tuning it for steerability. Experiments demonstrate effective steering across Diffusion Policy, $π_{0.5}$, and FastWAM. We further use steered rollouts for online policy improvement through DAgger. After 2-3 rounds of iteration, average success rates increase by 15.5\% for $π_{0.5}$ across three tasks and by 21.3\% across three policies(Diffusion Policy, $π_{0.5}$, FastWAM) on the Insert Bread task, while average human intervention counts decrease by 44.0\% (2.86 to 1.60) and 81.9\% (2.60 to 0.47), respectively.

0.28

YUBI-STAG: Contact and Semantic-Rich Alignment for VLAs via Automated Video-Language Grounding

cs.RO cs.CV ·2026-10-07 ·Masatoshi Tateno, Takehiko Ohkawa, Yueh-Hua Wu, Hanlong Li, Tatsuya Matsushima +2 more
code/data
Abstract · Project page: https://yubi-stag.airoa.io/

Vision-Language-Action (VLA) models acquire broad manipulation capabilities via large-scale pretraining, yet eliciting them through language requires fine-grained alignment between instructions and physical interactions. Existing robot demonstrations typically provide only coarse task descriptions, omitting how actions are executed, including which gripper acts, which object is contacted, and how it is grasped and moved. We introduce YUBI-STAG, a framework for Spatio-Temporal Annotation and Grounding that automatically enriches manipulation demonstrations with interaction-rich semantics to align pretrained VLAs with fine-grained manipulation language. Combining contact-object segmentation with vision-language models, YUBI-STAG annotates object identities, attributes and states, per-gripper actions, bimanual coordination, and spatially grounded interactions. To address YUBI-STAG's reliance on localized sequences and multi-stage VLM inference, we distill it into YUBI-VLM. YUBI-VLM directly recovers action structure and annotations from raw, unsegmented video in few inference calls and operates from wrist views alone. We evaluate both frameworks on YUBI-STAG-Bench across temporal, semantic, and spatial grounding tasks. YUBI-VLM retains much of YUBI-STAG's annotation accuracy with fewer inference calls and shorter runtime while generalizing to unseen manipulations. Finally, post-training VLA policies on these annotations aligns them with fine-grained language and contact-aware structure. Bimanual experiments demonstrate improved performance and instruction following, including control over object identity, acting gripper, target location, and spatial relations absent from original labels.

0.28

Unified Relative Orbital Element Framework for Eccentric Formation Flying

eess.SY math.DS ·2026-10-07 ·Yuji Takubo, Simone D'Amico
Abstract · 33 pages, 6 figures

Quasi-nonsingular relative orbital elements (qnsROE) are among the most widely used state representations for formation flying and rendezvous in near-circular inclined orbits. They admit closed-form perturbed state transition matrices (STMs) and a direct geometric mapping, which underpins eccentricity/inclination (E/I)-vector separation for passive safety. Existing eccentric extensions of the qnsROE resolve this geometry in the radial--tangential--normal frame, where a residual semimajor-axis difference makes the E/I separation margin phase dependent. This paper introduces eccentric ROE that reduce to the qnsROE as \(e\to0\) and whose first-order geometry is resolved in the velocity--normal--binormal (VNB) frame, where the secular drift is confined to the velocity direction. Their first-order position and velocity maps yield a closed-form, phase-independent binormal--normal separation bound with an explicit margin for semimajor-axis residuals. Closed-form Keplerian and secular-\(J_2\) STMs and a decoupled impulsive control input matrix are derived for the same elements, and the elements are recovered as the physically normalized integration constants of the Cartesian equations of relative motion in the VNB frame. Together, these results provide a unified ROE framework for formation flying in general elliptic orbits.

0.27

From Hitter to Player: Separating Motor Execution and Strategy in Competitive Humanoid Table Tennis

cs.RO ·2026-10-07 ·Fanrong Dong, Ziyi Sun, Ziqi Han, Junhan Sun, Lei Ye +4 more
code/data
Abstract · 8 pages, 7 figures. Project page: https://mufanmufan.github.io/hitter-to-player/

Competitive humanoid table tennis requires two distinct abilities: executing a desired return and deciding where to return the ball. We separate these two problems, with the ball landing target serving as an interface between motor control and strategy. (1) We first train a whole-body motor policy conditioned on the desired landing target, which maps the target to an executable humanoid stroke through a model-based strike planner and a learned controller. (2) We then improve the motor policy through self-play in closed-loop rallies, where the robot learns to handle incoming balls generated by another player rather than only prescribed incoming balls. At this stage, the landing targets are still sampled without considering the opponent, so the policy improves its execution without learning game strategy. (3) Finally, we freeze the entire motor stack and train a stroke-level strategic policy to choose landing targets through competitive self-play. Motor self-play improves both landing accuracy and legal-return rate during rallies. More importantly, with the entire motor stack fixed, changing only the target-selection policy leads to large differences in game outcomes: the learned strategy wins 89%, 82%, and 73% of points against center, random, and away-from-opponent baselines, respectively. The learned strategy also changes where it sends the ball according to the opponent's position and the previous target, showing that self-play learns more than simply aiming at fixed "good" locations. On hardware, the same frozen execution stack and strategic interface support complete 11-point games against a recreational human opponent, including multi-stroke rallies with both forehand and backhand returns.

0.27

TacHair: Tactile Contact-Distribution Guided Online Correction for Robotic Hair Stroking and Perception

cs.RO ·2026-10-07 ·Ruiyi Hu, Yongqiang Zhao, Daniel Bak, Yupeng Wang, Xuyang Zhang +1 more
code/data
Abstract

Hair stroking is common in daily grooming and personal care, and is also widely used in hair-product evaluation, motivating robots with similar physical interaction capabilities. Existing robotic hair-care and surface-following methods mainly rely on trajectory planning, compliance, force regulation, or tactile-conditioned policies, but deformable hair can remain in contact while gradually drifting across the end-effector, making local interaction difficult to regulate. We propose TacHair, a tactile contact-distribution guided online correction framework that represents high-resolution tactile observations as a spatial hair-contact distribution. A visuotactile imitation policy generates the nominal stroking motion, while a separately trained residual module corrects local contact deviations, separating task progression from contact recovery. We evaluate TacHair in 525 real-robot trials across five head geometries and three hair conditions. A successful stroke requires both sufficient task progression and contact maintenance; our method improves success from 42.9% to 62.3% and contact maintenance from 59.4% to 88.0% over the same visuotactile policy without correction. These results demonstrate spatial tactile contact distributions as an effective feedback representation for contact-preserving interaction with deformable and visually occluded surfaces. Demos, code, and datasets are available at https://tachair.github.io.

0.27

Juno: Taming Predictive Latents for Vision-Language-Action Models

cs.RO cs.CV ·2026-10-07 ·Yuchen Zhu, Chenyi Xu, Yulin Zhang, Gang Xu, Wentao Zhu
code/data
Abstract · Project Page: https://juno-policy.github.io/

Joint-embedding predictive architectures (JEPAs) predict masked or future observations in representation space, offering a natural source of predictive latents for vision-language-action (VLA) models. Yet making these latents useful across pretraining, policy learning, and deployment requires addressing three failures: mismatch with embodiment-specific control, interference with action learning, and teacher miscalibration under distribution shifts. We introduce Juno, a unified framework built around one action-conditioned JEPA that serves as a control-aligned representation backbone, a predictive teacher, and an adaptable dynamics model. During pretraining, we train it on embodiment-matched trajectories and use a dynamic CLS loss to transfer motion-weighted patch dynamics to a compact global state. During policy learning, we fuse current-frame JEPA patches into VLA perception and use a decoupled reasoning branch with separate transformation parameters to distill future latent states for action generation. During deployment, we adapt the world model on all observed transitions, including failed rollouts, freeze the adapted teacher, and re-align the policy on verified executions using LoRA adapters and a trainable action head, without expert corrections or task rewards. On SimplerEnv, Juno raises average success from $60.9\%$ to $68.5\%$ over Qwen3GR00T, the strongest baseline, and test-time adaptation further reaches $72.7\%$; on a real robot, it retains $70\%$--$75\%$ success under background, height, and object shifts where the base policy collapses to $0\%$.

0.27

Energy-Efficient Gait Adaptation via Hierarchical Reinforcement Learning for Quadrupedal Locomotion Across Diverse Terrains

cs.RO cs.LG ·2026-10-07 ·Ammar Issa, Anubhav Singh, Anton Tsaritsin, Sergey Kolyubin
Abstract · 9 pages. Submitted to IEEE ICRA 2027. Ammar Issa, Anubhav Singh, and Anton Tsaritsin contributed equally

While energy efficiency is a critical objective for legged-robot locomotion control, achieving low energy consumption while maintaining robust performance across different velocity ranges and terrain conditions remains a key challenge. This is particularly true for end-to-end RL policies, where gait generation, motion execution, and energy optimization are tightly coupled, leading to high sensitivity to reward design. In this work, we propose a hierarchical reinforcement learning (HRL) framework that separates a high-frequency policy for stable and robust joint-level motion execution from low-frequency gait adaptation that explicitly minimizes the cost of transport (CoT). The three-stage Isaac-based training procedure enables zero-shot sim-to-real transfer with improved tracking accuracy, robustness, and energy efficiency. The learned hierarchy exhibits automatic speed-dependent gait adaptation, transitioning from pacing at low speeds to trotting at higher speeds. We validate the proposed approach in simulation against representative single-policy and hierarchical locomotion baselines, demonstrating reduced CoT over a broad range of commanded velocities, while maintaining robust locomotion across flat, uneven rough, and inclined terrains. We further demonstrate its practical feasibility through zero-shot deployment on a physical Unitree AlienGo quadruped.

0.27

Towards Accurate End-Effector Localization for UMI-Style Robotic Manipulation Teaching

cs.RO ·2026-10-07 ·Junjie Zhang, Deteng Zhang, Zhisong Xu, Bo Sun, Liuyang Li +2 more
Abstract · 9 pages, 8 figures, 5 tables. Project page: https://mild-web.github.io

Robot demonstration learning requires accurate and temporally complete end-effector localization during close-range manipulation and camera occlusion. Existing SLAM benchmarks emphasize navigation motions, whereas manipulation datasets prioritize policy learning over localization evaluation. We introduce MILD, a Manipulation-Interface Localization Dataset with real-world and simulation sequences. The real-world subset provides 86 sensor sequences from Insta360 X5 and Insight9 across 15 repeated tabletop tasks, calibration assets, and a per-execution robot end-effector reference trajectory. The simulation subset, MILD-Sim, extends task coverage in Isaac Sim for controlled manipulation-replay studies. Benchmarking visual-inertial and fiducial-aided systems on instrumented real-world recordings reveals large differences in both TCP-relative trajectory error and temporal coverage, even under the same nominal task. To support marker-augmented teaching workspaces without a pre-surveyed fiducial map, we present AprilVINS, which combines fisheye visual-inertial estimation with sequence-local AprilTag geometry and separates prior admission from guarded export of the jointly optimized state. On Insta360 AprilTag4 recordings, AprilVINS(full) under a unified protocol with sequence-specific profiles reaches millimeter-level SE(3)-aligned TCP-relative APE RMSE with high time completion and lower reported error than the tested routes under their respective protocols, whereas fisheye VIO without tag factors remains at centimeter scale. Ablations separate accuracy from exportability, and a MILD-Sim replay study provides task-specific tolerance references for interpreting those error magnitudes. Together, MILD and AprilVINS provide a diagnostic benchmarking framework for UMI-style demonstration collection. Code, datasets, and evaluation manifests will be released upon acceptance.

0.26

Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment

cs.RO cs.LG ·2026-10-07 ·Xiao Zhang, Yuxin Chen, Zhixuan Liang, Guojian Zhan, Chenran Li +3 more
hot direction
Abstract

When diffusion policies were first introduced, they were expected to recover multi-modal action distributions. However, we find this expectation does not always hold, as diffusion policies often collapse to a single modality even when we guarantee the balance of dataset modalities and exact within-batch symmetry. Our analysis indicates that independent action-noise pairing contributes to this failure by increasing mixing and crossing among diffusion paths, which can produce averaged denoising responses and suppress modality-specific behavior. This issue is especially severe in robot planning, where action spaces are dense and low-dimensional, significantly increasing such mixing and crossing. To alleviate this problem, we propose Immiscible Diffusion Policy, a label-free training-time add-on to diffusion policy that uses action-noise assignment to preserve relatively distinct noise-to-action routes without modifying the policy architecture or inference procedure. Across five simulated and two real-world humanoid manipulation tasks spanning state, RGB, and point-cloud observations, our method significantly improves the policy's preservation of action modalities while maintaining strong task performance. It increases the proportion of the non-dominant modality by 6.0x-14.6x across three two-modality tasks and recovers demonstrated modalities that are entirely absent from vanilla policy rollouts on both four-modality tasks. These results demonstrate that Immiscible Diffusion Policy provides a simple yet robust approach to preserving action multi-modality in general robot learning tasks.

0.26

Continuous-Time Critic as a Lyapunov Function: Projected Adaptation and Cart-Pole Stabilization

eess.SY math.DSmath.OC ·2026-10-07 ·Pavel Osinenko
theory
Abstract

We study adaptive stabilization of known control-affine systems by treating the current parameterized critic itself as a Lyapunov candidate. Its total derivative contains a parameter-velocity term that can compensate unfavorable state evolution. A projection onto the intersection of a bounded admissible velocity set and one decay halfspace enforces parameter invariance and Lyapunov decrease simultaneously. We give an exact pointwise feasibility test, a conditional state-stability theorem, and a counterexample showing why a positive critic family containing a stabilizing member does not imply persistent feasibility. A convex endpoint constraint provides a distinct sampled implementation. A generic periodic polynomial-factor critic with a local Riccati anchor is evaluated on a force-limited cart-pole without mechanical-energy features, replay training, waypoints, or controller switching. The frozen controller passes stabilization and numerical decay checks on 600 sampled initial states across two draw sets; 400 matched continuous trajectories also pass. Independent reintegration audits and a conservative local feasibility certificate are distinguished from a global or machine-arithmetic guarantee. Compared with nonlinear model-predictive control, the measured implementation uses less CPU but has higher cost, settling time, and cart travel. The results support a conditional adaptive stabilization mechanism.

0.26

Temporal Visuo-Tactile Learning for Dexterous Grasp Stability

cs.RO cs.CVcs.LG ·2026-10-07 ·Ken Nakahara, Aleksei Buvailik, Prokhor Kotov, Roberto Calandra
code/data
Abstract · 12 Pages. Website: https://lasr-lab.github.io/dexterous-grasp-stability/

Humans can grasp everyday objects with almost perfect success rates using fingertip tactile feedback, yet much of the robotic grasping literature emphasizes vision-based grasp selection with parallel grippers. In this work, we systematically investigate how high-resolution, dynamic tactile sensing contributes to grasp stability prediction and model-guided grasping in dexterous robotic hands. To this end, we collected a dataset of 10,000 grasp trials across 200 objects using a multi-fingered robotic hand equipped with four Digit 360 tactile sensors, recording external vision, proprioception, and tactile streams throughout each grasp. With this dataset, we trained end-to-end temporal multimodal models to predict post-lift stability from pre-lift grasp observations and compared sensing modalities and encoding backbones. Experimental results and controlled input ablations show that incorporating touch, and particularly high-resolution, dynamic touch, improves grasp stability prediction. Finally, we deployed the learned predictor as an online stability gate on the real robot, where visuo-tactile model-guided regrasping improved the success rate among executed lifts by 10.5 percentage points over a non-tactile gate. These results show how rich fingertip sensing and expressive temporal models that capture the dynamics of touch can support learned grasping with multi-fingered hands without explicit contact or force modeling, providing a scalable data-driven path from tactile experience toward stable dexterous manipulation. The dataset is publicly available at https://lasr-lab.github.io/dexterous-grasp-stability/.

0.26

Skill-SLM: Agent Skill-driven Small Language Models for Reliable Robot Operation

cs.RO ·2026-10-07 ·Wenhao Wang, Yanyan Li, Jiawei Yuan
Abstract

Small language models (SLMs) have been increasingly adopted for onboard robot operation because they enable intelligent decision-making. However, existing approaches are mainly distillation-oriented and rely on enumerating representative task-solution pairs. This makes dataset construction difficult and limits generalization to diverse robot tasks whose possible forms grow rapidly. This paper proposes Skill-SLM, a framework that reformulates SLM-driven robot operation as a task-decomposition and skill-composition problem. Given a natural language task instruction, Skill-SLM decomposes the task into subtasks, selects appropriate skills from the skill library, and orchestrates the selected skills into executable robot operations. First, to support the skill-driven workflow, we propose a novel robot operational skill aware context-free grammar (CFG) to extract the skills required to accomplish tasks and build the skill library accordingly. Then, we configure LLM teachers to induce and synthesize training datasets for the SLMs, enabling SLMs to decompose tasks and orchestrate skills reliably. Additionally, we employ a progressive skill orchestration strategy to improve the reliability of skill implementation and overall robot operation. Experiments on UAV operation tasks indicate that Skill-SLM substantially outperforms distillation-oriented baselines, especially on unseen tasks that require generalization of capabilities. Additional experiments on ground vehicle tasks further demonstrate that Skill-SLM can be applied to different robot platforms.

0.26

Transition Path Sampling Using Koopman Operators and Exit-Time Optimal Control

eess.SY cs.LGstat.ML ·2026-10-07 ·Boya Hou, Shane Wang, Siddharth Ambekar, Maxim Raginsky, Olgica Milenkovic
Abstract · 30 pages, 6 figures

Sampling transitions between metastable states is a central problem in dynamical systems theory and molecular dynamics in particular. A key challenge is the existence of high free-energy barriers that separate the states, making transitions extremely rare. Recent machine learning-based methods cast transition path sampling (TPS) as an optimal stochastic control (OSC) problem over a fixed time horizon, and parameterize the drift bias via a neural network trained by simulation-in-the-loop, requiring repeated biased rollouts. To address computational and performance guarantee issues of these models, we propose a new approach for the problem based on Koopman operators. Because Koopman operators are linear, their leading eigenfunctions reveal the metastable sets and provide an estimate of the committor function with no transition path information required. Furthermore, we formulate TPS as an OSC problem up to an exit time. Our time horizon is the first hitting time of the target set, and our running cost penalizes time spent in nonreactive regions by encoding the estimated committor function. We derive the optimal controller in closed form and approximate it in a reproducing kernel Hilbert space (RKHS). This reduces the problem of constructing the optimal controller to solving a single equality-constrained quadratic program, whose solution can be characterized by a linear Karush-Kuhn-Tucker (KKT) system. On the two-channel double well and alanine dipeptide, our controller increases the fraction of trajectories reaching the target from 0% to 99.8% within 1000 steps, and from 0% to 93% within 1ps, respectively.

0.26

Semantic-Aware Predictive Mapping for Exploration and Navigation

cs.RO ·2026-10-07 ·Kenneth J. K. Ong, William W. J. Teo
published
Abstract · Accepted at IEEE TENCON 2026

Predictive mapping can support robotic exploration and navigation by estimating unseen geometric layouts from partial occupancy observations. However, occupancy-only representations may fail to distinguish semantically different structures with similar geometry. This is particularly relevant for indoor doors, which may appear as occupied cells like walls but indicate possible connected rooms or corridors beyond the observed region. This work investigates whether semantic door cues improve predictive geometric occupancy mapping around such ambiguous regions. We modify a subset of the CogniPlan dataset by inserting door-induced ambiguities into partial occupancy maps while keeping the ground-truth layouts unchanged. We compare a geometry-only control model with a semantic-cued model trained on the same modified dataset, where the semantic-cued model receives an additional door channel. Evaluation uses L1 error, F1 score, and Intersection over Union (IoU) over both the full map and a 10-pixel door-region mask. Full-map performance remains broadly similar between models, but localized door-region results show a clear qualitative improvement: L1 decreases from 0.004342 to 0.000025, while F1 and IoU improve from 0.031311 and 0.015905 to 1.000000 and 1.000000, respectively. These results suggest that semantic cues can improve predictive occupancy completion in regions where geometric observations alone are ambiguous.

0.25

Enhancing Robotic Perception and Adaptability through Sensor Fusion and Origami-Inspired Designs

cs.RO ·2026-10-07 ·Namai Chandra, Jaison Jose, Kavi Arya, Shivaram Kalyanakrishnan
Abstract · 11 pages, 11 figures, 3 tables

Compact mobile robots must recover scene geometry under changing lighting and surface texture while working within tight payload and cost limits. We present a compact mobile robot that uses origami-inspired wheels for locomotion and active control of its sensing geometry. As the wheels move between terrain-adaptive configurations, the changing chassis pitch sweeps a 2D LiDAR through intermediate elevations; held wheel positions provide a chosen viewing angle. An IMU accounts for chassis attitude, and a fusion node projects LiDAR returns into the RGB-D depth stream supplied to RTAB-Map. The arrangement uses the wheel actuation already present on a sub-300 USD, sub-2 kg prototype to extend the scanner's viewing geometry. We assess depth fusion in a textureless indoor corridor and an outdoor sunlit area, with three runs per sensor configuration in each setting. Mean full-frame invalid-depth fractions fell from 21% to 11% indoors and from 48% to 18% outdoors. The prototype combines improved depth coverage with a continuously adjustable LiDAR viewpoint using the same actuation that reconfigures its wheels.

0.25

Effects of Residual Chirp Rate on DAFT-s-AFDM Uplink Performance

eess.SP ·2026-10-07 ·Henrique L. Senger, Gustavo P. Gonçalves, Bruno S. Chang, Hyeon Seok Rou, Giuseppe Thadeu Freitas de Abreu
Abstract

Affine frequency division multiplexing (AFDM) is robust to doubly dispersive channels in high-mobility 6G systems but, like orthogonal frequency division multiplexing (OFDM), suffers from a high peak-to-average power ratio (PAPR). Spreading with a discrete affine Fourier transform (DAFT) yields DAFT-s-AFDM, whose precoder and modulator chirps can be chosen independently. For a single user with zero-offset localized allocation, we show that these two chirps merge into a single residual chirp whose rate, with the other two chirp rates fixed, alone determines their effect. At the transmitter, this rate defines a circulant spreading kernel whose fourth moment quantifies the envelope fluctuation; symmetries reduce its characterization to rates in [0, 1/4]. At a linear minimum mean-square error (MMSE) receiver, the residual chirp redistributes the mean- square error across symbols without changing its total; under a Gaussian approximation, a more uniform per-symbol error lowers the uncoded bit error rate (BER). In uncoded QPSK simulations, a zero rate yields the lowest PAPR, whereas a generic rate reduces the signal-to-noise ratio required for a BER of $10^{-3}$ by 4.4 dB at a 3.33 dB PAPR cost (exceedance probability $10^{-3}$);a rate of 1/3 captures 2.8 dB of this gain for only 0.88 dB of that cost.

0.25

Sparsity in control problems: Recent trends and open challenges

eess.SP math.OCeess.SY ·2026-10-07 ·Luca Ballotta, Geethu Joseph, Waseem Abbas, Sophie M. Fosson, Chandra R. Murthy +4 more
survey
Abstract · Submitted to Annual Reviews in Control. FIx author name

As in all fields of engineering, control engineers constantly face a dichotomy between achieving high performance and the limited availability of resources for sensing, actuation, and control. This tension often translates into formalizing design problems with sparsity constraints; that is, the design involves decision variables whose number, spatial reach, temporal spread, or other physical expressions are much smaller than in the ideal scenario of unconstrained optimization of a task-related performance metric. Even though this fundamental limitation underlines a plethora of methodological approaches and application domains, the control literature is fragmented into several lines of work tailored to specific scenarios. This survey fills the gap by bringing together several lines of work on handling sparsity in relevant control problems such as actuator scheduling, minimal-energy controllability, maximum hands-off control, and online convex optimization, and weaving a red thread through theoretical foundations and algorithms. Beyond establishing a coherent, structured narrative of problems and solutions for handling sparsity in control scenarios, we discuss open areas and relevant directions for future research, motivated by the growing complexity of control systems and their relations with key domains such as pervasive computing, distributed learning, cyberphysical security, and ecological footprint.

0.25

ActiveReg: Information-Driven Active Regional Probing for Partial-to-Full Bone Registration

cs.RO ·2026-10-07 ·Tiancheng Li, Yingyu Wang, Peter Walker, Liang Zhao, Shoudong Huang
Abstract · Preprint

Accurate bone registration is essential for orthopedic surgical navigation and robotic assistance. In a common workflow, the surgeon needs to identify and probe a number of prescribed locations on the exposed bone surface to obtain data points for registration, which can be difficult and time-consuming under limited surgical exposure. Reducing the number of required points while providing clear probing guidance would ease the surgeon's acquisition task intra-operatively. We present ActiveReg, a closed-loop framework that recommends probing regions rather than individual points, allowing flexibility in the exact contact location. A D-optimal planner that takes into account the already acquired points and the uncertainty of current registration information helps achieve accurate registration with fewer points. An online assessment combines pre-update innovation consistency with uncertainty at predefined surgical task locations improves the reliability of completion decision, without requiring ground truth. Simulations across four anatomical scenarios demonstrated that ActiveReg can achieve competitive accuracy compared with baseline registration methods while using substantially fewer acquired points. Real phantom experiments with optical and electromagnetic tracking demonstrated comparable mean target registration error to the Gradient-SDF framework, using only 24-28 acquired points instead of 771-1263.

0.25

MultiFly: A Real-World Multimodal Aerial Dataset with Annotation-Efficient Label Transfer and Cross-Modal Semantic Consistency

cs.RO cs.CV ·2026-10-07 ·Markus Gross, Andreas Greiner, Taehyoung Kim, Sivasubiramaniam Subbiah, Tomaž Cotič +9 more
benchmarkcode/data
Abstract

We introduce MultiFly, a real-world, low-altitude UAV dataset for semantic perception across RGB, thermal, LiDAR, and radar modalities. MultiFly provides 17,272 synchronized samples from four suburban scenes with frame-wise annotations for 15 semantic classes, together with calibration and GNSS-RTK/IMU measurements. To avoid costly and inconsistent modality-specific annotation, we propagate labels from only 115 manually annotated RGB images through shared geometric representations to all four modalities. This approach generates semantic labels for 17,157 additional RGB images, 17,272 thermal images, 840M LiDAR points, and 3.4M radar points. Transferred annotations achieve 89.93% average agreement with held-out manual annotations, and 90.94% average semantic consistency across all six modality pairs. We further establish semantic segmentation benchmarks for all four modalities, revealing distinct architectural behavior for dense LiDAR and sparse radar data. Taken together, MultiFly provides a scalable foundation for multimodal aerial perception and, to the best of our knowledge, the first public real-world low-altitude aerial benchmark that combines consistent frame-wise semantic annotations for RGB, thermal, LiDAR, and radar. Data at https://github.com/markus-42/multifly.

0.25

Moving Horizon Estimation of Hybrid Systems Using Hybrid Zonotopes and Mixed-Integer Quadratic Programs

eess.SY ·2026-10-07 ·Jonah J. Glunt, Herschel C. Pangborn
Abstract

Moving horizon estimation (MHE) is an estimation technique that formulates an optimization problem over a past horizon to find the states that require the minimum disturbances and noise to be feasible with the dynamics and measurements. For hybrid systems with both continuous and discrete states, this yields a mixed-integer program for the MHE, which can be difficult both to formulate and solve online under computational limitations. This paper shows how the hybrid zonotope representation for mixed-integer sets can be used to efficiently construct the feasible space for the program using reachability-based calculations for piecewise affine systems, without requiring conversion to a mixed logical dynamical system. A method for lifting the state space to explicitly estimate the dynamic mode is also presented. The proposed approach is evaluated in two numerical examples. Comparison against a standard mixed logical dynamical systems approach shows the proposed method can be solved orders of magnitude faster for one of the example problems.

0.25

Higher-Order Morphology Priors for Quadruped Reinforcement Learning Under Actuator Degradation

cs.RO cs.LG ·2026-10-07 ·Derek You, Zafir Shamsi, Keqin Wang, Christine Allen-Blanchette
published
Abstract · Accepted to IROS Workshop BLPC 2026

Actuator degradation turns quadruped locomotion into a coordination problem requiring joints to compensate for lost actuation. Prior work suggests that morphology-aware graph policies improve learning and generalization under body perturbations. We ask whether these benefits can be strengthened by explicitly modeling higher-order mechanical structure. We represent the Unitree Go1 as a cell complex with limb- and body-level rank-2 cells and apply Hodge-based message passing. Under degradation training, the node-edge-face Hodge actor achieves the highest return on unseen actuator degradations, with higher survival and lower velocity-tracking error. These results support higher-order morphology as a useful inductive bias for whole-body compensation under actuator degradation.

0.24

Noise-Induced Navigation in Non-convex Domains and Compact Manifolds

cs.RO math.OCeess.SY ·2026-10-07 ·Karthik Elamvazhuthi
Abstract

In this note, we study the problem of designing a feedback law that globally steers a system to a prescribed target configuration. Even if the system is fully actuated, topological obstructions generally prevent the existence of globally asymptotically stabilizing continuous feedback laws. We revisit this problem in a stochastic setting by allowing noise to enter through the control channels. Using a criterion for asymptotic stability in the large that combines local Lyapunov stability with positive recurrence, we constructively show that one can construct elementary feedback laws that achieve global asymptotic stability in the large, of the target equilibrium point in connected Euclidean domains with obstacles and manifolds without boundary. For Euclidean domains with obstacles, we also show that the method extends naturally to the problem of finding the minimizer of a strongly convex function with non-convex constraints. Numerical experiments illustrate the effectiveness of the approach for Euclidean domains with circular obstacles and the two dimensional sphere. Additionally, we study the role noise strength when there is a non-convex obstacle, in which case the system might show metastability.

0.24

Evaluating the Transfer of Co-Evolved Communication from 2D to 3D Simulation

cs.RO cs.NE ·2026-10-07 ·Fernando Montes-Gonzalez
Abstract · 14 pages, 5 figures. Preprint

This work examines the transfer of a co-evolved communication mechanism between two robotic agents from a discrete two-dimensional (2D) simulator to a three-dimensional simulator with real physics (3D). The study focuses on whether a communication mechanism co-evolved in a 2D environment retains its functional role after transfer to a 3D physics-based simulator. To support this analysis, the effects of the episode time budget, the social cue, and the asymmetry between the two co-evolved roles were examined. The results indicate that the success rate increased approximately linearly with the evaluated time budgets, with no evidence of a plateau between 2,000 and 6,000 physics steps, suggesting that evaluations based on shorter episodes may underestimate the performance of the trained controllers. In both simulators, the social cue functioned primarily as a jam- assistance mechanism rather than as a navigation guide, although with a more pronounced effect in 2D. Analysis of eight independent evolutionary runs revealed a consistent direction of asymmetry, although its magnitude varied across runs. Controlling the processing order between agents allowed us to rule out an artifact of the physics engine. Finally, the results are discussed in terms of the factors that may contribute to the remaining performance gap observed after transfer.

0.24

Robust Prediction of Internal Wave-Affected Multi-Scale Sound Speed Distribution Using Lightweight Kolmogorov-Arnold Networks with Hybrid Basis Functions

eess.SP cs.SD ·2026-10-07 ·Wei Huang, Junpeng Lu, Tianhe Xu, Hao Zhang, Feng Yin
Abstract

The underwater sound speed distribution directly governs acoustic propagation paths, rendering it critically important for underwater acoustic communication and target localization. Conventional sound speed profile (SSP) prediction methods provide a good way to estimate the underwater sound speed distribution without on-site data measurement, thus breaking through the coverage area constraints of sonar observation equipment and making the model universal in most marine areas. However, underwater sound speed exhibits multi-scale variations, such as diurnal, quarterly, and intermittent fluctuations caused by ocean processes such as internal waves. This makes it difficult for the fixed structure models in existing methods to have good generalization ability for multi-scale sound speed distribution patterns. To tackle this problem, we proposed a lightweight hybrid basis-function empowered Kolmogorov-Arnold network (LHBF-KAN) model for multi-scale sound speed prediction. We aim to construct a multi-branch representation layer in which different basis functions respond to distinct temporal patterns, from slowly varying background trends to rapid fluctuations induced by dynamic ocean processes, allowing the model to naturally accommodate the inherently multi-scale evolution of sound speed at different depths. To prevent the multi-branch structure from increasing model size, a pruning strategy is further introduced to suppress branches with consistently low contribution during training, yielding a compact architecture, suitable for deployment on resource constrained underwater platforms.

0.23

Not All Uncertainty Matters: Simulation-in-the-Loop Fast-Slow Reasoning for Decision-Critical Autonomous Driving System

cs.RO cs.AI ·2026-10-07 ·Jiayi Chen, Shuai Wang, Guangxu Zhu, Derrick Wing Kwan Ng, Chengzhong Xu +1 more
Abstract

Large vision-language models (VLMs) provide powerful open-world perception and reasoning for autonomous driving, but their high computational cost and inference latency make continuous cloud-side use impractical. This motivates fast--slow collaboration, where efficient onboard modules handle real-time perception and control while cloud models provide high-level reasoning only when needed. The key challenge is deciding when cloud reasoning should influence time-critical driving decisions. Existing methods often rely on perception uncertainty, heuristic triggers, or resource-driven policies, without assessing whether resolving an uncertainty will improve planning. We propose \textbf{SIGMA}, a simulation-in-the-loop framework for task-oriented fast--slow collaboration. SIGMA embeds the planner into uncertainty assessment and evaluates how plausible scene realizations under semantic and geometric uncertainty affect feasible trajectories and planning cost. Based on these outcomes, it estimates the expected reduction in planning cost from resolving uncertainty. We further introduce expected planning gain (EPG), a decision-level metric for cloud invocation, cloud-guidance integration, and request prioritization under deadline and resource constraints. Experiments in CARLA show that SIGMA reduces unnecessary cloud interactions while improving planning, efficiency, and navigation success in static and dynamic obstacle scenarios. Compared with fixed-period collaboration, SIGMA reduces unnecessary cloud interactions by 50\%, improves navigation success by more than 6\%, and cuts finish time by up to 26.2\% in dynamic scenarios.

0.23

Compute-Constrained Safety Filters with Neuromorphic Event Triggering

eess.SY math.OC ·2026-10-07 ·Tochukwu E. Ogri, Opeyemi Owolabi, Luke Fina, Christopher Petersen, Rushikesh Kamalapurkar
Abstract · 11 pages, 4 figures, submitted to 2027 IEEE American Control Conference

Safety filters that rely on control barrier function (CBF) quadratic programs (QPs) are typically analyzed under the assumption that the QP is solved instantaneously, although nonzero computation time can render a computed input unsafe by the time it is applied. This paper develops a dual leaky integrate-and-fire (LIF) event-triggered safety filter for linear systems that accounts for this delay. Two LIF states monitor a shifted CBF residual and the deviation of the held input from the nominal feedback law. A computation is initiated when either LIF state reaches its threshold; the previous input remains applied until the prescribed application time, and any solution returned earlier is stored until then. A closed-form charging profile relates the safety-LIF threshold to residual-growth bounds, ensuring sufficient residual margin over the application delay. Under the stated assumptions, the resulting hybrid system guarantees forward invariance, completeness, and non-Zeno execution, with positive lower bounds on computation-initiation and input-application times. A spacecraft terminal-approach simulation demonstrates safety with fewer optimizer calls than a delay-tightened periodic filter.

0.23

Restorative Control and the Limits of Safe Model Discrimination

eess.SY math.OC ·2026-10-07 ·Micheal Myers
theory
Abstract · Submitted to the 2027 American Control Conference (ACC 2027)

Restorative control seeks lasting recovery through an intervention that ends in finite time. We formulate this objective as a reach-avoid problem in which the input must bring the system safely into a desired basin of attraction before withdrawal. Two candidates share bistable baseline dynamics but accumulate irreversible damage through different allocations of drive-generated exposure. We derive a lower bound on the exposure required for basin transfer and a sufficient construction for safe restoration. When the lower bound exceeds the summed tolerances, every common open-loop input that restores either model causes failure under at least one. With sufficient tolerance, proportional allocation restores both without identification. For feedback from output samples with Gaussian noise, we bound safe-restoration probability in terms of failure risk under the alternative model. A numerical example illustrates the certificates under two sensing regimes.

0.23

Tactile Reconstruction of Contact Task Frames and Forces for Hybrid Force/Motion Control

cs.RO ·2026-10-07 ·Antonio Rapuano, Simone Orelli, Barbara Bazzana, Antonio Franchi, Alessandro De Luca
Abstract · 14 pages, 9 figures, 7 tables. Video: https://youtube.com/playlist?list=PLejKMZmW8AvI. Dataset: https://doi.org/10.5281/zenodo.23209032

Hybrid force/motion control requires knowledge of the interaction force and of a task frame defining the force- and motion-controlled directions. These quantities are usually obtained from force/torque sensing or model-based residuals, often assuming also a nominal environment model. This work addresses the online estimation of the contact force and a possibly time-varying task frame using only soft optical tactile sensing, under the assumption of locally planar contact with a negligible contact moment. The proposed method maps a single image of the deformed elastomer of a soft optical tactile sensor to observable contact variables: indentation depth, two surface-to-sensor tilt angles, and 3D contact force, each with a per-sample uncertainty estimate. The mapping is learned through a self-labeling acquisition procedure, in which a manipulator imposes controlled contacts while an auxiliary Force/Torque sensor is used offline to provide ground-truth labels. The tactile measurement is then fused with robot proprioceptive data in an Extended Kalman Filter, producing a continuously updated estimate of the contact task frame and of the interaction force. Control experiments with a DigiTac sensor mounted on a UR10 manipulator demonstrate closed-loop contact force regulation against a flat rigid board in linear and angular motion by a human operator, with touch as the only exteroceptive feedback.

0.23

Robotic Boomerang Throwing via Model-Based Release Design

cs.RO ·2026-10-07 ·Yang Liu, Colin Jones, Aude Billard
code/data
Abstract · Project page: https://robot-boomerang.github.io

Throwing objects that generate aerodynamic lift can greatly extend robot throwing beyond ballistic flight. A returning boomerang is a challenging example because its flight depends strongly on the release velocity, attitude, and spin, while robotic manipulators cannot readily reproduce the rapid motions used in human throwing. We present a model-based framework for robotic boomerang throwing centered on the release state. We identify the boomerang flight dynamics in stages to predict how flight changes across design variations. To systematically design the robot throwing motion, we screen candidate parameters according to how strongly and consistently they control release spin under uncertain contact conditions. These models are then used to design the throwing motion and boomerang for a 6-DoF manipulator with limited joint speeds. To our knowledge, this is the first robotic manipulator to generate a returning boomerang flight. In the demonstrated returning trial, the boomerang is released at 51 rad/s (8.1 rev/s), reaches 2.03 m from the robot base, and returns to touch down 0.31 m from the base. The successful release differs significantly from the measured human throws, showing that a robot need not imitate human throwing motion to achieve a returning flight. The project page is available at https://robot-boomerang.github.io

0.23

ΔWAM: Distilling Action Tangent Fields into World Action Models

cs.RO ·2026-10-07 ·Ke Wu, Hanwen Huang, Bo Gu, Kaizhao Zhang, Xiangting Meng +4 more
Abstract · 9 pages, 4 figures

World Action Models (WAM) improve robot policies by augmenting sparse action supervision with dense future prediction. However, much of the predictable future is dominated by appearance and scene persistence rather than action-dependent dynamics. We observe that several recent WAM designs, including optical flow, motion-centric representations, and latent actions, can be understood from a common perspective in which world supervision becomes more efficient as it contains a higher proportion of action-relevant variation. Based on this insight, we introduce Action Tangent Fields, which reformulate world supervision through a local Taylor expansion of how actions induce changes in future dynamics. We represent future dynamics in Residual-VAE space, where the future latent remains recoverable from the current latent and its residual, and use a strong action-conditioned world model (ACWM) to probe the local correspondence between action variations and residual-world variations. This local first-order structure is distilled into the WAM to guide its denoising supervision toward dynamics that are more tightly coupled to action, rather than merely predictable from appearance. Across LIBERO-Plus, RoboTwin, and RoboTwin2.0-Plus, our method consistently improves robustness to lighting, background, camera, layout, and other environmental perturbations. Despite using no large-scale embodied pretraining, it achieves stronger robustness under several distribution shifts than pretrained policies. We further distill multi-step VideoDiT denoising into a single step for efficient inference. Our results suggest that effective WAM supervision should remain information-rich while concentrating its predictive capacity on the directions along which actions change the future.

0.22

High-Fidelity Baseline Design and Station Keeping Analyses for Earth-Moon Vertical Orbits

eess.SY ·2026-10-07 ·Derek Phung, Yuri Shimane
Abstract · 20 pages, 11 figures, 7 tables. Submitted to the 2027 AAS/AIAA Space Flight Mechanics Meeting

Earth-Moon vertical orbits are periodic solutions of the circular restricted three-body problem (CR3BP) with substantial out-of-plane motion, considered for high-latitude lunar observation and cislunar navigation. This work develops an end-to-end framework for high-fidelity reference generation under rotating-frame geometry constraints and receding-horizon model predictive control. A representative $L_2$ vertical orbit yields a 5.5-year reference with negligible deterministic correction and remains maintainable in five-year Monte Carlo simulations under realistic uncertainties. The workflow produces locally optimal solutions across 90 consecutive daily start epochs and 200 catalog cases. Finite-Time Lyapunov Exponent along the reference and stable/unstable CR3BP directions interpret station-keeping behavior.

0.22

Learning Stability of Replay-Based Co-Optimization for Transmission Expansion under Strategic Bidding

eess.SY ·2026-10-07 ·Tomonari Kanazawa, Hikaru Hoshino, Eiko Furutani
Abstract

This paper investigates the behavior of learning-based co-optimization for transmission expansion under strategic bidding in electricity markets. In this framework, transmission capacities are updated while market participants simultaneously learn their bidding strategies through deep reinforcement learning, resulting in coupled and non-stationary learning dynamics. We show that transient policy degradation of bidding agents can generate inconsistent cost-capacity samples, which bias the transmission-capacity update and prevent the co-optimization process from converging to the desired solution. To mitigate this, replay-based capacity updates with nearest-neighbor filtering are introduced to exclude inconsistent samples from the update data. We then analyze a new oscillatory behavior that appears when the replay memory is enlarged. Numerical results on the IEEE 30-bus system demonstrate that enlarged replay memories improve robustness against transient policy degradation but can introduce a temporal lag, leading to oscillations unless the capacity-update learning rate is appropriately reduced. These results reveal the trade-off between replay-memory size and learning rate and provide practical guidelines for stable co-optimization.

0.22

Flexible Analog Self-Interference Cancellation for Waveform-Robust Monostatic ISAC

eess.SP ·2026-10-07 ·Anh Tuyen Le, Xiaojing Huang, J. Andrew Zhang, Peiyuan Qin, Le Chung Tran +2 more
Abstract

Monostatic integrated sensing and communication (ISAC) requires in-band full-duplex operation, where self-interference cancellation (SIC) must suppress the strong transmit leakage while preserving weak sensing echoes. In the analog domain, this task is complicated by transmitter impairments, which generate conjugate and nonlinear SI components, and by waveforms with widely different statistics. This paper develops a flexible analog self-interference cancellation (FASIC) framework for monostatic ISAC based on the widely linear normalized least-mean-square (WL-NLMS) algorithm. FASIC generates the cancellation signal in digital baseband, subtracts it in the radio-frequency domain, and enables flexible sensing-range control via a configurable cancellation span. Its normalized adaptation remains stable across waveforms with substantially different peak-to-average power ratios (PAPRs), for which fixed-step adaptation can become unstable. We further relate transmit-waveform statistics to analog SIC performance and predict the waveform-dependent cancellation floor. Simulations validate the analysis and demonstrate stable analog SIC of approximately $30$--$34$ dB, with the low-PAPR waveforms achieving a $3.3$ dB deeper floor owing to their smaller high-order amplitude moments. Near-field static reflections within the cancellation span are suppressed while delayed and Doppler-shifted echoes are preserved. These results establish FASIC as a waveform-robust analog SIC solution and identify waveform statistics as a key consideration for sixth-generation (6G) waveform selection.

0.22

Learning Unknown Constraints without Unsafe Data via Optimality and Counterfactual Regularization

cs.RO cs.LG ·2026-10-07 ·Zhouyu Zhang, Chih-Yuan Chiu, Glen Chou
Abstract

Learning from demonstrations (LfD) provides a framework for inferring unknown constraints from locally optimal, constraint-satisfying expert behavior. Existing approaches largely fall into two paradigms, constrained inverse optimal control (CIOC) and inverse constrained reinforcement learning (ICRL). CIOC exploits optimality conditions such as the Karush--Kuhn--Tucker (KKT) conditions but typically assumes known dynamics and structured constraint representations. Meanwhile, ICRL accommodates complex unknown constraints and unknown transition dynamics but often requires extensive online exploration, during which unsafe constraint violations may occur. In this work, we introduce Counterfactual KKT (CF-KKT), a constraint learning framework that leverages learned dynamics and locally optimal demonstrations to recover unknown constraints without requiring known dynamics or additional risky exploration, thereby combining the data efficiency and safety advantages of CIOC with the flexibility of ICRL. First, we use a locally learned differentiable dynamics model to impose KKT-inspired optimality conditions directly on the demonstrations. Second, we use the learned dynamics to generate reward-improving counterfactual behaviors near the demonstrations, revealing behaviors that would be preferable in the absence of the unknown constraint and thus providing synthetic infeasible data. When the constraint parameterization is known, the same learned-dynamics framework enables direct CIOC-based parameter recovery, and we characterize its sensitivity to dynamics misspecification. Across high-dimensional robotic control tasks, our approach learns neural constraint representations with improved safety and data efficiency relative to state-of-the-art offline ICRL baselines.

0.22

Sparse Feature Policy Unlearning Mitigates State Hallucination in Vision-Language-Action Models

cs.RO cs.AIcs.LG ·2026-10-07 ·Jiho Lee, Jeongeun Park, Heayoun Choi, Taekyung Kim, Eunwoo Kim
Abstract

Vision-Language-Action (VLA) models have shown strong generalization in robotic manipulation by leveraging rich representations from pretrained vision-language models. However, their deployment in real-world environments remains limited by recurring unreliable behaviors. In this work, we study state hallucination, a recurring failure pattern in which a VLA continues acting as if an unrealized robot-object state had been achieved. Our analyses find that state hallucination coincides with weakened attention to task-relevant visual regions, and a mechanistic interpretation via sparse autoencoders reveals that hallucination-associated sparse features are activated when these failures occur. Based on this analysis, we propose SOUL (Sparse feature pOlicy UnLearning), which selectively unlearns policy knowledge associated with state hallucination behaviors, where sparse features identified from hallucination failures and successful behaviors serve as explicit forgetting and retention targets, respectively. Experiments across VLA architectures in simulated and real-world environments show that our method substantially reduces hallucinated failures and improves task success without substantially compromising the existing manipulation capabilities. These results suggest that interpretable feature analysis provides a practical basis for selectively modifying undesirable knowledge in robot policies.

0.21

RobotAPO: Adversarial Physics Preference Optimization for Robotic Manipulation Video Generation

cs.RO ·2026-10-07 ·Kerui Li, Zhe Jing, Chenyi Huang, Xiaofeng Wang, Zheng Zhu +2 more
Abstract

Robotic manipulation videos are increasingly used as visual plans for embodied agents, but optimizing purely for visual plausibility often fails to capture the fragile physical manifold of real-world interactions. Even minor physics-violating errors at the interaction boundary, such as interpenetration or premature object motion, can completely invalidate the inferred timing and pose needed for downstream execution. Because standard supervised fine-tuning lacks the direct pressure to penalize these localized failures, we introduce AgiBot-PhysPref. This rigorously curated 10,000-sample preference dataset isolates condition-matched physics violations, turning the generator's own failure distribution into a foundational signal for physical consistency. Building upon this, we propose RobotAPO, an adversarial physics preference optimization framework operating in the continuous flow-matching denoising space. To prevent the policy from merely memorizing static curated failures, RobotAPO employs a lightweight adversarial counterfactual proposer that learns a condition-dependent, physical-failure-biased direction in denoising space. This encourages the model to explore and better respect the physical interaction boundary, all while maintaining a pure prompt-and-reference inference interface without requiring external structural conditioning. Comprehensive evaluations demonstrate that explicitly correcting these localized physics violations improves downstream robot execution from generated videos. On held-out AgiBot conditions, RobotAPO outperforms the strongest controlled internal baseline in physical consistency by 6.8% hard score and 10.0% soft score. Crucially, in real-robot replay, it translates these physical-consistency gains into a 37.4% relative improvement in task success over the strongest controlled internal baseline.

0.21

Masked Generative Motion Planning with Geometry-Guided Token Search

cs.RO cs.AI ·2026-10-07 ·Lipeng Zhuang, Yingdong Ru, Shiyu Fan, Edmond S. L. Ho, Gerardo Aragon Camarasa +1 more
Abstract

Generative motion planners typically use learned trajectory priors for initial generation, while leaving test-time repair to local continuous refinement. We introduce Masked Generative Motion Planning (MGMP), which extends the learned prior from efficient parallel generation to structural repair. A masked generative transformer generates discrete trajectory candidates in parallel, and Geometry-Guided Token Search (GGTS) uses scene geometry to target where to edit and which prior-supported alternatives to evaluate. This turns refinement into an efficient search over discrete motion alternatives, enabling route-level restructuring beyond local trajectory deformation. MGMP achieves 96% success on Ring Maze and 82% repair success on Controlled Route Invalidation on Kuka, exceeding the strongest external baselines by 23 and 25 percentage points, respectively. It further generalizes to unseen layouts, additional obstacles, unseen geometries, single- and dual-arm planning, and real-world Baxter tasks.

0.20

RoboQuest: Generalist Physical Agents that Search, Inspect and Test

cs.RO cs.AIcs.CV ·2026-10-07 ·Liu Renhang, Navonil Majumder, Tej Deep Pala, Soujanya Poria
benchmark
Abstract

Recent advances in multimodal foundation models have made them capable generalist physical agents for a range of manipulation tasks. However, successful operation in an unfamiliar environment may require an agent to seek task-relevant information through interaction when it is absent from the observations: it may need to determine where a relevant object is, inspect an unobserved property, or discover the effect of an unfamiliar tool. We thus introduce RoboQuest, a benchmark for goal-directed embodied exploration, where agents must actively acquire task-relevant information through physical interaction, use the resulting evidence to adapt subsequent actions, and autonomously decide when to commit to task completion. RoboQuest comprises ten mobile manipulation tasks centered on three forms of uncertainty: search, manipulation-based inspection, and interactive testing. We evaluate five frontier multimodal agents through a common visuomotor interface, as well as a $π_{0.5}$ policy fine-tuned on the full-episode demonstrations we release. The best agent succeeds in only 23\% of the episodes, and the fine-tuned policy almost never succeeds. Isolated tests of the execution skills the tasks are built from, with the hidden information supplied, show that the agents can carry out most of the required actions, and our failure analysis attributes only a minority of the failures to execution. Our failure analysis further finds that the agents often stop exploring too early as they make decisions before observing the required evidence for task completion. We also find that agents rarely prevent or repair the disturbances caused by their exploration. Moreover, learning by trial and error remains difficult for most models.

0.20

Borrowed Eyes: Markerless Nano-UAV Flight with an Active Quadruped Observer

cs.RO ·2026-10-07 ·Alejandro Lorite Mora, Dimitrios Arapis, Andrés Faíña
Abstract · 8 pages, 7 figures, 2 tables. Submitted to ICRA 2027

Nano unmanned aerial vehicles (nano-UAVs) can navigate confined spaces that larger robots cannot, but their payload capacity severely limits the sensors and compute available for self-localization in global navigation satellite system (GNSS)-denied environments. We present a vision-based system that localizes a nano-UAV from a quadruped robot with an arm-mounted camera. The quadruped tracks the drone, estimates its position in its own coordinate system using segmentation masks and depth, and transmits that position over a real-time radio link. To ensure continuous tracking, we developed a perception-aware nonlinear Model Predictive Controller (NMPC) that dynamically adjusts the quadruped's body and arm to maximize the drone's visibility, treating observation reliability as a primary control objective. Relying solely on this external estimate, the nano-UAV executed predefined trajectories from takeoff to landing with a median 3D localization error of 59 mm. The system demonstrates robustness against short visual occlusions by smoothly transitioning to onboard inertial flight when line-of-sight is temporarily lost. Ultimately, this framework allows a quadruped to offload the localization burden of a nano-UAV, enabling inspection of complex spaces that neither robot could navigate alone.

0.20

AeroEval: Staged Program and Execution Validation for AI-Generated Drone Missions

cs.RO cs.DC ·2026-10-07 ·Kautuk Astu, Naina Rabha, Yogesh Simmhan
Abstract

Large Language Models (LLMs) can generate drone programs from natural-language mission descriptions, but syntactically valid programs may still violate user intent, environmental constraints, and mission-level behavior. This problem is pronounced in cyber-physical applications, where correctness depends on the interaction among generated code, mobile sensing, environmental geometry, event-driven analytics, and physical execution. Existing drone code-generation systems primarily use prompt guardrails or simulator outcomes and provide limited failure localization. We present AeroEval, an agent-assisted middleware for staged validation of AI-generated drone missions. AeroEval combines deterministic program analysis with context-grounded LLM agents. It first validates program syntax, platform API usage, and mission intent, and then evaluates the realized behavior using execution trajectories, mission requirements, and environmental context. Each stage returns structured failure information for iterative regeneration. In our evaluation using 20 navigation tasks and five analytical mission types over AirSim and Gazebo simulators, AeroEval improves navigation success from 55% to 95%. In a stagewise ablation study, our Code and Trajectory Validators by themselves achieve mean run-level success rates of 44% and 56%, respectively, while the full AeroEval pipeline achieves 88%; the stages detect complementary failures in program structure, API usage, mission intent, obstacle avoidance, altitude, coverage, and event-driven transitions and the guided regeneration corrects for them. Across the main analytics missions, AeroEval increases aggregate run-level success from 34% for one-shot AeroGen to 88% within the regeneration budget. These results demonstrate the benefit of combining program-level and execution-grounded agentic validation for AI-generated drone applications in the evaluated environment.

0.20

EnergyNet in Practice: Long Grid, Short Grids and Mobility-Based Energy Peering

eess.SY ·2026-10-07 ·Jonas Birgersson, Max Collins, Marc A. Weiss, Jimmy Chen, Daniel Kammen +8 more
position
Abstract

Electrification is planned as a network problem: more demand, more upstream capacity, local resources waiting for reinforcement. We argue that a material part of that bill, in money and time, follows from architecture, not electrification itself, and propose the Short Grid as the deployment unit: a locally controlled building, block or campus whose demand, generation, storage and local market operate behind Energy Routers enforcing an approved connection envelope to the Long Grid. Because the envelope governs exchange, not installed capacity, local resources can grow and many sites can build in parallel at the pace of PV and EV installation, not grid reinforcement. PV and EV batteries dominate today, but the architecture is general: any generation or storage, fixed or mobile, fits behind the same boundary and interfaces. EnergyNet Operators coordinate exchange among Short Grids and with the Long Grid; protection stays local; Vehicle-to-EnergyNet (V2EN) and Mobility-Based Energy Peering (MEP) make mobility-financed batteries accessible within and between them; priority-based operation protects critical services in shortage. We test it in models, not yet the field: a ten-building neighbourhood in Lund, Sweden; a bus-linked microgrid pair at Stanford and Half Moon Bay; and a proposed 20-site network in Rockaway, New York. Households and services are the largest demand category in Sweden, Denmark, Norway, California and New York (484 TWh/yr); halving their grid imports would be about 242 TWh. Europe's network investment to 2040 is estimated above EUR 1.2 trillion; modelling shows it depends on system design. Negagrid value, following the negawatt, counts network investment avoided, downsized or deferred, net of local substitute costs. "Don't pay twice": build coordinated local capability first; reinforce the Long Grid for what remains.

0.19

When Listening Becomes Easier: Scrubbing Visual Cues for Shortcut-Free VLAs

cs.RO cs.CVcs.LG ·2026-10-07 ·Jasper Gerigk, Kenzo Aspuru-Takata, Chin-Hsuan Wu, Mohammad Mohammadi, Shuhong Zheng +1 more
Abstract

Shortcut learning is a prevalent issue in robot learning. The limited diversity of robot demonstration datasets can mislead policies into exploiting spurious correlations between tasks and irrelevant features, such as viewpoint or background. Collecting sufficiently diverse robot demonstrations is costly and inefficient, motivating algorithmic alternatives. We focus on vision-language-action (VLA) models and discover that different vision-language model backbones exhibit substantially different levels of susceptibility to visual shortcut learning. We find that model behavior correlates with our proposed representation-level metric, action margin, which requires no policy rollouts. Visual shortcuts consistently enter action representations in early layers, with models differing in the extent to which later layers correct them by incorporating language information. To boost models' attention to language, we introduce task scrubbing, a new domain-adversarial training method that decreases models' likelihood of using visual shortcuts and improves VLAs' generalization. Experiments in both simulation and the real world across multiple VLAs and visual cues show that task scrubbing improves out-of-distribution robustness and often eliminates visual shortcut learning.

0.18

Hierarchical Defense in Leader-Defender-Attacker Games: Nested Equilibria and Receding-Horizon Control

eess.SY ·2026-10-07 ·Honglei Guo, Hanfang Zhang, Guochuan Zhang, Yuhan Zhao
Abstract

We formulate a discrete-time leader-defender-attacker game that combines hierarchical interaction within a protective group with competition against an external attacker. The leader seeks to reach a prescribed demand point while avoiding capture, whereas the defender seeks to intercept the attacker while remaining near the leader. To capture their distinct objectives and asymmetric interactions, we introduce a nested equilibrium concept combining a leader-defender Stackelberg relationship with a Nash-type best-response condition for the attacker. For linear dynamics and quadratic objectives, we derive affine response laws and a backward recursion, together with sufficient conditions for unique stagewise solutions. We further express arrival, capture, and interception conditions as quadratic inequalities in the initial joint state, characterizing winning regions under the resulting policy. A receding-horizon algorithm implements the computed control laws using updated state information. Numerical experiments over randomized initial configurations examine the influence of the prediction horizon and compare the proposed method with an independent-control baseline, illustrating differences in game outcomes and the agents' realized costs.

0.18

Factorized Tactile Representation and Control for Sim-to-Real Manipulation

cs.RO ·2026-10-07 ·Siqi Shang, Bianca Aumann, Tye Brady, Joshua Migdal, Taskin Padir
Abstract · 8 pages, 7 figures

Tactile sim-to-real learning must bridge simulated contact and device-specific sensor responses while preserving information needed for control. We propose a factorized tactile representation and control framework that maps normal force and contact patch to an effective contact response recoverable from sensor readings. The response is separated into contact geometry, force distribution, and temporal contact change, with representation-specific encoding and randomization. A Tactile Gated Policy preserves these representations separately through control and operates over all mask configurations without retraining. We evaluate the approach through response reconstruction, spatial alignment, force regulation, and contact-rich adversarial peg insertion in simulation and the real world, enabling the utility and transfer reliability of different tactile representations to be assessed independently. The approach achieves <1 mm contact localization, 1.69 N force-tracking error on unseen geometries, and a 35% improvement in real-world adversarial peg insertion over the unfactorized response, with different tactile representations benefiting different interactions.

0.18

HuMBLE: Human Motion-Driven Behavior Learning for Embodied Locomotion

cs.RO ·2026-10-07 ·Mike Zhang, Dongho Kang, Kevin Bergamin, Nicola Burger, Robin Deits +17 more
Abstract

Despite recent advances in humanoid locomotion, controllers optimized for command tracking and robustness tend to produce mechanical gaits, whereas controllers tied to human motion data often fail to generalize to commands outside the data distribution. This work introduces a learning framework that balances these competing objectives to synthesize real-time steerable, robust, and biomimetic locomotion policies from human data. Using an in-house curated locomotion dataset covering diverse speeds and directions, we first learn a natural locomotion prior policy through a teacher-student distillation process. Specifically, we train a full-body reference-conditioned policy with Reinforcement Learning (RL), then distill it into a lightweight prior policy conditioned solely on proprioception and a planar torso-velocity steering command. Next, we fine-tune the prior policy with multi-task RL to expand command coverage and robustness beyond the data distribution, pairing a goal-conditioned task that tracks arbitrary commands with a reference-guided task that tracks the human data as an explicit style regularizer. We validate our framework on three humanoid robots: the Boston Dynamics Atlas R1, Atlas D1, and Unitree G1. Experimental results demonstrate robust performance across real-world scenarios, including direct user-controlled locomotion in indoor and outdoor environments, and integration as the locomotion layer within hierarchical control stacks. Benchmarks against Tabula Rasa RL policies trained without human data and ablation studies confirm that our framework yields a lightweight, deployable policy that reconstructs coordinated whole-body behavior from a steering command, retaining the human gait characteristics while remaining robust and fully steerable.

0.18

Decoding Neural Population Dynamics through Robotic Analog

cs.RO ·2026-10-07 ·Wenhui Chen, Jiyue Tao, Yitao Cheng, Yutong Shi, Feitian Zhang +2 more
Abstract

Animal evidence shows that precise voluntary movements arise from rotational neural population dynamics in motor cortex, but their physical effects remain unknown. We developed a robotic analog of biological motor systems with artificial muscles, multimodal sensors, and a neural network controller trained via reinforcement learning. The robotic analog exhibited accurate movements, robustness to damage, and neural population dynamics akin to animals. This task-driven, embodied model illuminates the causal link between neural population dynamics and motor outcomes. We discovered that neural rotations generate oscillatory maneuvers orthogonal to the reaching direction, optimizing trajectory adjustments, which is confirmed by primate neural data. The model also revealed counterintuitive neural energy principles under sensor and motor redundancies, and striking Eureka moments during motor learning, bridging biological and artificial systems. These findings provide new perspectives on how neural dynamics contribute to accurate and flexible movement, inspiring future intelligent robots with animal-like mobility.

0.18

Do Vision-Language-Action Models Understand Instructions? A Mechanistic Interpretability Study on Language Grounding

cs.RO ·2026-10-07 ·Theodor Wulff, Angelo Cangelosi
benchmark
Abstract

Vision-Language-Action models are designed to generalise across environments and task descriptions, raising the question of whether their action generation actually depends on the language instruction, or whether they largely rely on visual cues and superficial correlations. Robustness to variance in the visual and linguistic observation space is critical for real-world deployment, yet VLAs lack explicit grounding modules and instead rely on the intrinsic language grounding capabilities of their Vision-Language model backbones. For this reason, we conduct a controlled mechanistic interpretability study on the language grounding capabilities of two state-of-the-art Vision-Language-Action models, $π_{0.5}$ and GR00T N1.7, by applying activation and attribution patching to the residual stream of the action generation modules. We systematically corrupt the task instruction of input samples of the LIBERO benchmark following five strategies: synonym replacement, semantic scaling, directional corruption, random object substitution, and empty string. Our experiments find that both models are comparatively insensitive to abstract rephrasing and to referencing non-existent objects, but react strongly to empty task descriptions and, especially, to directional language. During action generation, this sensitivity is concentrated in different loci for each model: mainly in the early, periodic cross-attention layers for GR00T N1.7, versus distributed across the earliest and selected later layers for $π_{0.5}$. For GR00T N1.7, directional perturbations drive some of the largest causal effects while leaving the internal representational geometry comparatively unchanged, a dissociation we do not observe clearly for $π_{0.5}$. Finally, the reliability of attribution patching is model-dependent: it closely tracks activation patching for GR00T N1.7 but not for $π_{0.5}$.

0.18

Do My Samples Help? Compute Budget Allocation for Real-World Sampling-Based MPC in Contact-Rich Manipulation

cs.RO ·2026-10-07 ·Magnus Dierking, João Carvalho, Vince Kurtz, Jan Peters, Nils Dengler +1 more
Abstract

Sampling-based model predictive control (SMPC) is increasingly practical for contact-rich manipulation because GPU-based physics simulators enable real-time evaluation of large numbers of non-differentiable contact rollouts. However, this computation budget is finite and can be spent in competing ways, including denser control exploration, representing dynamics or state uncertainty, replanning more frequently, or aligning the simulation better with the physical state. Yet, the resulting trade-offs are not well understood in real-world sim-to-real control. In this work, we address this question through a systematic study of how to allocate a fixed computation budget for SMPC in contact-rich manipulation. To enable controlled comparisons, we develop a GPU-based framework named Budget-Allocating Model Predictive Control (BAMPC), which runs on MuJoCo Warp and exposes dynamics uncertainty, state uncertainty, and temporal alignment as independent dimensions of the computation budget while using the same control stack in simulation and on a real robot. Across three contact-rich manipulation tasks, we find that the best budget allocation depends on the dominant sim-to-real error, with dynamics randomization not universally the most effective one.

0.18

Hall Effect-Based Tactile Force Detection Sensor for Robot-Assisted Minimally Invasive Surgery

cs.RO ·2026-10-07 ·Charles DeLorey, Michael Brockdorff, Adam Metcalf, Benjamin Calmé, Pietro Valdastri +1 more
Abstract · 9 pages, 12 figures

Robotic minimally invasive surgery has revolutionized surgical practice. However, as the surgeons are mechanically separated from the surgical tools, loss of tactile feedback occurs. We present a compact force sensor with three, millimeter scale sensing elements (taxels) that can be affixed to the grasping face of a robotic surgical tool. The sensor features a miniaturized design using Hall effect sensors and magnets embedded in a deformable elastomer contact layer. Sensor calibration is achieved with a 6 degree of freedom (DoF) parallel robot. Each taxel can detect applied normal and shear forces with average RMSE of 2.133 kPa. The calibrated sensor was validated on phantom and ex vivo porcine tissue, demonstrating tactile sensing during retraction as well as slip. This work introduces a small-scale Hall effect based tactile sensor, enabling sensing for force-sensitive robotic surgical applications.

0.17

Trajectory Sensitivity-Based Dynamic Assessment of an AI Data Center Under Contingencies

eess.SY ·2026-10-07 ·Arnav Bagga, Saskia A. Putri, Abhishek Banerjee, Xiaofan Wu
Abstract · 7 pages, 5 figures

Artificial Intelligence (AI) data centers are emerging as converter-dominated, parameter-rich electrical systems whose dynamic response depends on fast workload-driven demand variations, tightly regulated data-hall converters, and coordinated on-site sources. Assessing such systems by repeatedly solving nonlinear time-domain models for every outage, load realization, and parameter perturbation is computationally expensive and provides limited insight into response sensitivity. This paper presents a trajectory sensitivity (TS)-based dynamic assessment framework for a self-contained AI data center supplied by a 2 kV DC microgrid with a grid-tied interface. The model includes time-varying data-hall load profiles, a voltage source converter (VSC), battery energy storage systems (BESS) units, supercapacitors, and synchronous-generation-side sources. For each disturbance, firstly, TS are computed along the nonlinear disturbance trajectory, capturing the local effect of source trips, data-hall outages, converter controls, and load ramps. Later, the computed TS data is used to estimate system trajectories under simultaneous parameter perturbations. Validation against full nonlinear ODE simulations with $\pm 15\%$ perturbations shows close agreement. In addition, the TS-based approach reduces computational time of dynamic-response evaluations from seconds to milliseconds, demonstrating its potential as a scalable screening tool for AI data center dynamics.

0.17

AirGroundVLN: A Large-Scale Benchmark for Goal-Oriented Air-Ground Collaborative Vision-and-Language Navigation

cs.RO ·2026-10-07 ·Zhenxuan Zeng, Qingle Wu, Wei Suo, Maojia Wu, Bairong Zhang +2 more
benchmark
Abstract

Goal-oriented Vision-and-Language Navigation (VLN) requires agents to locate and reach targets described in natural language without prescribed routes. Air--ground collaboration is valuable for tasks requiring both wide-area search and fine-grained localization. However, systematic study of goal-oriented air--ground collaborative VLN remains limited by the lack of large-scale, diverse benchmarks and two core challenges: 1) substantial differences between aerial and ground views, together with useful observations becoming unavailable as navigation proceeds, make it difficult to maintain spatially consistent context across platforms and over time; and 2) asymmetric spatial observability makes ground perception locally detailed but spatially limited and aerial perception broad but locally coarse, limiting the reliability of single-platform planning. To address these limitations, we introduce AirGroundVLN, a benchmark containing 10,281 navigation episodes and 955 target instances across 19 Unreal Engine environments, with seen/unseen splits and an aerial-visibility protocol for systematic evaluation. Alongside the benchmark, we propose AG-CoNAV, a trainable reference framework comprising two key components: Spatiotemporally Anchored Collaborative Memory (SACM) and Aerial-Guided Regional-to-Local Planning (AGRLP). SACM maintains and retrieves spatially consistent historical context across aerial and ground observations. Meanwhile, AGRLP combines regional aerial guidance with fine-grained ground navigation. Extensive experiments demonstrate the effectiveness of AG-CoNAV and establish AirGroundVLN as a comprehensive benchmark for future exploration.

0.15

NAViLoss: An Underwater Navigation-Aware Dual-Residual Objective for Physics-Consistent Learning

cs.RO cs.LG ·2026-10-07 ·Arup Kumar Sahoo, Itzik Klein
Abstract · 26 pages, 7 figures

Autonomous underwater vehicles (AUVs) commonly rely on inertial navigation systems (INS) aided by Doppler velocity logs (DVLs) for reliable underwater navigation. Accurate DVL velocity estimation is therefore essential for successful operation. Recent learning-based methods have demonstrated improved DVL velocity estimation, particularly under degraded measurement conditions. However, their training objectives typically rely on conventional regression losses that are highly sensitive to large residuals and corrupted observations. Additionally, they do not explicitly account for the physical consistency and measurement uncertainty associated with the underlying sensing process. To address these limitations, this paper introduces navigation-aware loss (NAViLoss), a robust and uncertainty-aware objective function for learning-based AUV velocity estimation. NAViLoss jointly penalizes the velocity-estimation residual in the navigation-state domain and the beam-consistency residual in the DVL measurement domain. Its bounded formulation limits the influence of large residuals, while an adaptive mechanism regulates the uncertainty in beam geometry. Furthermore, NAViLoss is integrated with a DeepONet architecture to form a novel NAVi-DeepONet model for seamless estimation of an underwater vehicle's velocity. Lastly, our model is evaluated using approximately 10,000m of semi-synthetic AUV experimental data collected during multiple real-world sea trials. Experimental results demonstrate a 44% improvement in velocity-estimation accuracy compared with conventional and learning-based baselines. These results demonstrate the effectiveness of navigation-aware and uncertainty-adaptive loss design for robust learning-based underwater velocity estimation.

0.15

People & Society 39

Procedural Generation of Conceptual 3D Ship Internal Arrangements for Design-Space Exploration

cs.CE ·2026-10-07 ·Kirsten Odendaal, Roy de Winter
code/data
Abstract · 17 pages, 7 figures, 5 appendices

Ship internal arrangement is a complex spatial design problem in which competing functions within a fixed volume affect spatial relationships, mass distribution, trim, and stability. Exploring these trade-offs requires many layouts in a common format, yet public general-arrangement drawings are scarce, inconsistent, and two-dimensional. This study develops a common framework for generating conceptual 3D internal arrangements across six vessel families: bulk carriers, tankers, container vessels, offshore support vessels, patrol vessels, and yachts. Each arrangement uses paired voxel grids and topological graphs for spatial and connectivity analysis. The framework is evaluated through three experiments: (i) comparison of 50,000 generated arrangements with 23 reconstructed real references, (ii) envelope-constrained reconstruction of one reference per family, and (iii) response of generated layouts to requested function volumes. Generated design spaces broadly align with references within each family's sampling range. Searching alternative spatial subdivisions improves reconstruction across all families, with voxel accuracy reaching 75.5%. Fuel and ballast allocations respond near-proportionally to requested volume, while other functions depend on subdivision capacity, compartment eligibility, and discrete assignment. A fixed-hull offshore support vessel application shows that the same hull and requested programme can produce different adjacency and trim outcomes, with five nondominated concepts among 119 of 500 alternatives meeting stability and programme criteria. The results support layout design-space exploration across vessel families and changing requirements, with spatial organisation remaining important beyond volume compliance. The generator and conceptual general arrangement dataset are released publicly for reproducible analysis, benchmarking, and learning-based design studies.

0.41

CrossWeave: Bridging Perspectives Across Online Communities with a Dual-Pane Design

cs.SI cs.HCcs.IR ·2026-10-07 ·Fei Fang, Reva Hirave, William Jurayj, Yuqi Li, Brian Lu +17 more
published
Abstract · CSCW 2026 + small improvements

Social media systems typically display conversations among already familiar contributors, which can be predictable and one-sided. In civic discourse, this design narrows discussion, reinforces divides, and distorts the perception of public opinion. To encourage cross-community engagement, we present CrossWeave, an AI-powered bridging system that augments the standard social media feed. As the user reads a post, CrossWeave surfaces diverse relevant posts from other threads in a side pane and highlights the connections. Users are invited to venture out of their echo chamber, explore a broader range of views and arguments, and ``click across'' to engage with their authors. When they do, CrossWeave facilitates constructive posting, not only by showcasing relevant past content but also by simulating possible reactions as the user drafts a post.

0.41

Some Twitch chats exhibit multifractal characteristics resembling stream-of-consciousness narrative

cs.SI nlin.AOphysics.data-anstat.AP ·2026-10-07 ·Marcin Ból, Stanisław Drożdż, Robert Kluszczyński, Jarosław Kwapień, Marcin Wątorek
published
Abstract · accepted for publication in CHAOS: An Interdisciplinary Journal of Nonlinear Science

Live-streaming platforms generate high-frequency records of collective human activity in which many loosely coordinated users react in real time to shared external stimuli and to one another. We investigate whether the temporal organization of Twitch chat messages exhibits multifractal signatures characteristic of complex dynamical systems. Selected high-activity chats were converted into time series by measuring the length of each consecutive message in words, with platform emotes treated as individual tokens. The resulting signals display heterogeneous fluctuations, bursty organization, and heavy-tailed distributions that, in several cases, are better captured by q-Weibull forms than by conventional light-tailed models. Autocorrelation analysis reveals persistent temporal dependence over broad ranges of message lags, indicating that the variability is not produced by independent message generation alone. MFDFA shows that some chats possess well-developed multifractal spectra and generalized Hurst exponents strongly dependent on moment order, whereas others exhibit weak scaling, approximately monofractal behavior, or no well-defined scaling regime. Surrogate tests indicate that temporal correlations are essential for the observed multifractality, while heavy tails enhance its apparent strength. Under particular dynamical conditions, Twitch chat may therefore form an emergent collective signal whose fragmented, reactive, associative, and multiscale organization is structurally reminiscent of stream-of-consciousness narrative. This expression is used strictly as a structural and dynamical analogy, not as a claim of a unified collective mind or of semantic, cognitive, or phenomenological equivalence with literary stream-of-consciousness writing. The results position live-stream chat dynamics within the broader physics of complex, intermittent, human-generated signals.

0.39

Active Inference for Interaction-Mediated Control of a High-Dimensional Robotic Arm

cs.HC ·2026-10-07 ·Fraser C. Paterson, Sebastian Stein, Markus Klar, John H. Williamson, Roderick Murray-Smith
Abstract · 21 pages. 4 figures. To be published in the Springer CCIS series as part of the 7th International Workshop on Active Inference (IWAI26)

We propose interaction-mediated control via Active Inference as a general architectural approach to high-dimensional control problems in Human--Computer Interaction (HCI). This architecture recasts user interaction as the provision of evidence about a latent task objective, rather than the direct specification of plant-control inputs. The mediating function is distributed between an interaction broker, which selects informative user queries and performs Bayesian inference over user preferences, and an Active Inference controller, which autonomously plans and acts under the resulting preference information to control the plant. This division of labour decouples the semantics of user interaction from those of low-level plant control. We instantiate the architecture in a simulated, multi-link robotic arm to perform a simultaneous whole-arm target-coverage task. A simulated user communicates exclusively through a clutch-style binary evaluative channel, without specifying joint-torque commands. Across increasing arm dimensionalities, the architecture achieves successful interaction-mediated control, although task success is lower than when the controller receives the true target preferences directly. Exact target-subset identification also remains imperfect, highlighting the distinction between preference inference and successful task completion. The experimental findings provide an initial computational demonstration of the proposed architecture under controlled, matched-model assumptions and motivate its further investigation in broader HCI applications.

0.34

A General Capacity Frontier of Complex Networks

cs.SI physics.soc-ph ·2026-10-07 ·Nicholas Kunz, H. Oliver Gao
Abstract · 38 pages, 4 figures, 9 supplementary tables, 2 supplementary notes

From cells to cities, ecosystems to economies, complex networks exhibit peak intensities that define their empirical limit. Estimating these limits traditionally requires domain-specific knowledge. We demonstrate that a wide range of complex networks share a capacity frontier that constrains these maxima. We formalize this regularity by proposing Structural Capacity Theory for Network Substrate Structured Systems. These systems couple a fixed network substrate with an observable dynamical process. The frontier is learned from a vector representation of the network substrate and modulated log-additively by a vector representation of the dynamical process. Frontiers trained on one set of domains reliably bound empirical maxima of entirely different ones without observing their peak intensities or simulating their dynamics. Independent learning paradigms converge on consistent upper bounds. Frontier performance is stable under perturbation, yet degrades under deliberate falsification, with no performance gain under ablation. These findings support the capacity frontier as a generalizable constraint that describes how network structure bounds dynamical processes across otherwise unrelated systems.

0.33

A dynamic pore-network model with discrete microbubble transport

cs.CE physics.flu-dyn ·2026-10-07 ·Kristoffer Skjelanger, Anna Mareike Kostelecky, Maziar Veyskarami, Timo Koch
unusual
Abstract · 32 pages (+11 in supplementary) and 20 figures (+18 in supplementary)

We develop a hybrid pore-network model (PNM) for discrete microbubble transport coupled to connected two-phase flow in porous media. We apply the model to simulate flow in the porous transport layer of a rotating proton exchange membrane (PEM) electrolyser. At sufficiently large apparent gravitational forces, bubbles growing at the catalyst layer can detach at sizes smaller than the characteristic pore size and move through liquid-filled pores via buoyancy and advection. The model couples a dynamic two-phase PNM to a Lagrangian bubble tracker. Pore-scale Stokes simulations in square throats provide closures for the bubble buoyancy velocity, the hydraulic resistance of bubble-occupied throats, and the reciprocal coupling between liquid flow and bubble motion. Verification tests assess conservation, pressure-flow coupling, bubble growth, advection, merging, trapping, and transfer to the connected gas phase. The network simulations show that increasing apparent gravitational forces shifts a larger fraction of the produced gas toward discrete-bubble transport. The increase mostly comes from a reduced detachment radius, which decreases bubble size and lets them move more freely through the network.

0.33

Many Brains, One Geometry: A Shared Visual-Semantic Space for Cross-Dataset fMRI Decoding

cs.HC q-bio.NCq-bio.QM ·2026-10-07 ·Moein Khajehnejad, Michelangelo Tronti, Forough Habibollahi, Tommaso Boccato, Matteo Ferrante +1 more
Abstract · 21 pages, 5 figures, 1 table, 3 supplementary figures, 2 supplementary tables

Visual decoding from fMRI is typically siloed by participant and experiment, obscuring whether heterogeneous neural measurements can be organized within a common computational geometry. Here we introduce BRAID-fMRI (Brain Representation Alignment across Individuals and Datasets), a shared CLIP-supervised decoding framework. BRAID-fMRI uses a single ROI-wise Transformer with optional participant conditioning across eight visual-fMRI datasets comprising 93 dataset-specific participant entries, 430,007 single-trial responses and 162,839 unique stimuli. Regional brain activity is aligned with 512-dimensional CLIP ViT-B/32 representations using a multi-positive contrastive objective that treats repeated stimuli across participants and datasets as positives. BRAID-fMRI supports retrieval across seven evaluation datasets. On eight matched participant entries, it achieves 35.0 +/- 11.1% Top-10 accuracy, exceeding the observed mean accuracy of the two evaluated baselines - the MindEye-style pooled-CLIP decoder (27.1 +/- 5.1%) and ridge regression (21.1 +/- 9.1%) - and attaining the highest observed accuracy for seven of eight entries. In separately trained participant-agnostic models, expanding the source pool increased target-dataset holdout accuracy by up to 92.1% relative to the initial source-training condition. The learned space preserves graded semantic structure, while ablations and saliency highlight ventral and early visual cortex and category-specific motion and attentional systems. These results support scalable cross-dataset decoding into a common CLIP-aligned space, with model sensitivity concentrated in ventral and early visual inputs.

0.32

When Exposure Is Not Attention: Auditing the Preference-Exposure-Consumption Gap in Personalized News Recommenders

cs.SI ·2026-10-07 ·Woojin Park
Abstract · 16 pages, 5 figures, 14 tables. Submitted to ICWSM 2027

Personalized news platforms are often evaluated as if stated preferences, logged recommendation exposure, and click consumption form a single coherent pipeline. Collapsing these layers can distort audit conclusions: a platform may appear more aligned or diverse than observed click consumption supports, which can misdirect diversity governance or algorithmic intervention. We introduce a reusable Preference-Exposure-Consumption (PEC) audit framework that separates stated preference, observed weighted profile state, logged recommendation exposure, app-surface pathways, and click consumption under explicit observability boundaries. Using six months of logs from a deployed mobile news application, we audit 1,583 user profiles, 95,143 logged recommendation items, and 17,512 click events. Each trace type contributes distinct information; none directly substitutes for another. Preference-consumption alignment exceeds chance-based null baselines but captures only part of users' top-consumed category set. Logged recommendation lists contain clicked articles more often than a date-matched candidate-pool baseline predicts (11.29% vs 9.13%; top-5 lift 1.47x), but many clicks arrive through other app surfaces. Raw exposure-consumption diversity gaps shrink under count matching, yet concentration mismatch persists in the audit-eligible cohort (HHI gap 0.082). Together, the results show that audit conclusions change depending on whether platforms measure stated preference, logged exposure, or click consumption.

0.31

Beyond Activation: Gaze Invocation with Visible Status for an Embodied AR Assistant in Co-Located Collaboration

cs.HC ·2026-10-07 ·Chenrui Ma, Yoshio Ishiguro, Qing Zhang
Abstract · 11 pages, 9 figures, 2 tables

Wake words face an inherent trade-off: higher sensitivity reduces missed commands but increases accidental activations. In co-located augmented reality (AR), the system must also determine whether the user is speaking to the assistant or to a nearby person. We introduce a new perspective on this problem: combining gaze-based address with an embodied assistant, visible listening status, and turn management across activation, continued interaction, and release. In our implementation, users look at an assistant anchored in the scene and see activation progress and listening status on its body. Sustained gaze opens a local interaction channel; speech and playback keep it open when attention returns to the task; and inactivity closes it. We evaluated this design with 25 participants. After selecting the assistant's placement and dwell duration, each participant worked with a partner to plan a trip while using the assistant. The study recorded 500 interaction outcomes, including 496 assistant-directed utterance attempts. Gaze acquisition completed before speech for 490 of these attempts (98.8\%): 467 began while the channel remained open, whereas 23 began after it had been released. The other 6 attempts began before acquisition completed. Among 102 reviewed attempts in which gaze left after acquisition but before speech, the configured retention policy kept the channel open for 79 and released it before 23. The results show that an embodied target with visible status can support clear entry into an assistant interaction while revealing a different problem at release: after users look back to their work, they may not see that the assistant has stopped listening. We contribute the implemented gaze-invocation design and design implications for communicating assistant state after visual attention moves elsewhere.

0.29

A Scale For Value Alignment In Human-AI Interaction

cs.HC ·2026-10-07 ·Lena Hegemann, Steeven Villa, Hyemin Bang, Mitchell L. Gordon, Antti Oulasvirta +2 more
Abstract

Value alignment is a central objective in AI and HCI research, yet no validated instrument measures how users perceive it. This gap hampers the comparison and accumulation of findings and limits the effectiveness of applications, where understanding users' viewpoints is critical. We construct and evaluate a 13-item psychometric scale that measures perceived value alignment across two components: value understanding and value manifestation. It is based on a large item pool drawn from prior empirical studies, filtered by experts, and finally assessed by users (N=607) across diverse AI scenarios. Confirmatory factor analysis on an independent sample (N=259) confirmed the two-factor structure and high internal consistency for both subscales. Using optimization, we also derived a 6-item short form for quick administration. The scale is a reliable measure, providing HCI researchers with a common evaluation metric across contexts.

0.27

An ecology of participation for fusion energy development

cs.CY physics.soc-phcs.HC ·2026-10-07 ·Aditi Verma, Katie Snyder, Andrea Morales Coto, Nathan Kawamoto, Daniel Hoover +8 more
position
Abstract

Decisions that will shape future fusion facilities, including the production of waste, the management of tritium, the achievement of safety, and impacts on land, water, and local communities, are increasingly becoming sociotechnical in nature. For prior energy technologies, fission among them, such decisions were made top-down and met sustained public opposition. Research shows such opposition is rarely a deficit of public understanding and is instead rooted in place-specific environmental, health, sociocultural, and economic concerns and in distrust of how technologies are selected and sited. We argue that fusion developers have an opportunity to design differently, and we propose an ecology of participation: multiple, complementary modalities of engagement sustained over time, replacing the one-off public consultations now typical of large infrastructure projects. We synthesize human- and environment-centered design frameworks and introduce a framework that reinterprets opposition as a divergence between communities and technology developers in values, norms, or design characteristics. We present six case studies from our work: participatory technology assessment focus groups, participatory design workshops, immersive virtual-reality models of a fusion facility, the Global Fusion Forum platform, the Imaginary Energies speculative design platform, and Heartbeat, an arts-based sound installation. Mapping these onto the values-norms-design framework shows how each interrogates or closes a different part of the expert-public gap, and exposes two limitations of fusion public engagement efforts writ large: developers are seldom asked to articulate their own values and norms, and the environment is represented only indirectly. We close with five guiding questions for fusion researchers building their own participatory engagements.

0.27

When Scientific Cognition Is No Longer Scarce

cs.CY cs.AI ·2026-10-07 ·Nathan DeBardeleben
Abstract · 14 pages, 1 figure

AI could change which parts of science impede progress. Consider a world in which machine systems are better, faster, and cheaper than people at most scientific work that can be done through a computer. Our question is what would limit science in that world. Literature synthesis, hypothesis generation, software development, simulation, and analysis could become abundant, while experiments, observations, well-supported conclusions, and accountable institutional au- thority remain scarce. Science would then be constrained by a different set of resources. In this paper, we call this change the scarcity inversion and consider four parts of it: selection, physical access, validation, and organizational choice. This change is arriving first in mathematics and coding/software/algorithm design, where the whole scientific loop can run inside computation. For national laboratories, the change could be striking. Their distinctive role is to turn abundant machine reasoning into trustworthy results by combining controlled experiments, protected data, expert judgment, and accountable authority. The practical question is how facilities, verification, provenance, resource allocation, and scientific governance should change when reasoning is plentiful and trustworthy evidence is scarce.

0.26

EEG and Eye-Tracking Evidence That AI Disclosure Shapes Face Evaluation

cs.HC cs.AI ·2026-10-07 ·Teodora Mitrevska, Luise Donat, Andreas Butz, Thomas Kosch, Abdallah El Ali +1 more
Abstract

AI-generated faces can be difficult to distinguish from real ones, leaving viewers to rely on source labels when judging an image. Yet prior work has made it difficult to separate the effects of what an image actually is from what viewers are told it is. We validated faces as AI-generated or human in an online study (N=169), then crossed actual source (AI, human) with label (none, Made with AI, Made by a human) in a lab study $N=30), recording event-related potentials (ERPs) and gaze. ERP responses were equivalent for AI-generated and real faces, but varied with the label: labels drew early attention (N2), while labels that conflicted with the face's actual source prompted re-evaluation of the face (P3). Affective processing and initial gaze orienting were unchanged, but labels altered visual exploration. We provide a validated stimulus set and evidence that attributed origin shapes face processing, with implications for disclosure design.

0.26

When My Skill Becomes Agent Skill: How Knowledge Workers Share Their Expertise with AI Systems

cs.HC ·2026-10-07 ·Tianqi Song, Zicheng Zhu, Hancheng Cao, Yi-Chieh Lee
Abstract

Organizations have long sought to make workers' expertise reusable by others. Agentic AI changes the nature of such reuse by enabling AI systems to act on workers' knowledge with limited human involvement. This raises the question of how this shift shapes workers' willingness to share their expertise. We conducted an experiment with knowledge workers who created materials incorporating domain knowledge and decided whether to authorize human or AI reuse. We find that AI and human reuse differ primarily in whether workers choose to share their knowledge, rather than in what they choose to share. Participants' reasoning shifts from prosocial considerations when sharing with humans toward concerns about loss of control, replacement, and downstream governance when sharing with AI. These findings suggest that AI reuse may intensify tensions between organizational knowledge reuse and contributors' interests. We discuss implications for workplace AI and knowledge management policies that preserve workers' rights and agency.

0.26

Empowering Users in Graph Rule Mining via Large Language Models

cs.HC ·2026-10-07 ·Francesco Cambria, Francesco Invernici, Andrea Colombo, Anna Bernasconi
Abstract

In the era of interconnected data, graphs have emerged as an effective abstraction for modeling complex systems in an intuitive format, especially with the rise of Property Graphs, which offer an intuitive and scalable way of navigating non-intuitive structures. In this context, graph mining techniques have been developed for testing complex graph-based rules, as the MINE GRAPH RULE operator, which, however, require users to have prior expertise both in graph theory and formal query language. In this work, we propose to bridge the gap between users and the graph-association rule-mining process by showing how Large Language Models (LLMs) can be easily prompted to formulate, refine, and interpret complex relational rules, directly producing MINE GRAPH RULE queries.

0.25

Healthy skepticism in AI: a data visualization research agenda

cs.HC cs.AI ·2026-10-07 ·G. Elisabeta Marai, Marc Baaden, Michael Behrisch, Michael Krone, Pere-Pau Vázquez
position
Abstract

Research in data visualization of artificial intelligence (AI) models has historically focused on enhancing trust through visual explanations of AI. The trustworthiness line of work was built at least partially on an assumption that humans were critical users unlikely to adopt AI technology. It is increasingly clear that human trust levels in AI span, in fact, a wide range from critical to over-reliant. There is an urgent need to support both trust and healthy skepticism in AI solutions. We argue that it is healthy for humans to adopt a skeptical view both on the results of AI models and on the use of such AI models. We share our thoughts on the rising phenomenon of over-reliance on AI models, the risks and opportunities in using AI models, and the role of data visualization in over-reliance situations where humans are not motivated to engage in critical thinking.

0.24

Guided Reflection for Personal Sleep Insight in Everyday Sleep Tracking

cs.HC ·2026-10-07 ·Bokyung Kim, Amama Mahmood, Honghao Zhao, Molly E. Atwood, Luis F. Buenaver +2 more
position
Abstract

Digital sleep technologies make tracking accessible, yet users often struggle to interpret what changes in their sleep mean. Behavioral sleep medicine addresses this through guided discovery, helping patients develop personal interpretations rather than simply receiving explanations. To bring this to everyday tracking, we present DREAM, an LLM-powered voice assistant that monitors conversational sleep diaries, selectively invites users to interpret meaningful changes, and uses their interpretation to tailor subsequent education. We co-designed DREAM with sleep specialists iteratively and evaluated it in a six-week field study (N=14) against a generic-education control. DREAM participants reported greater personal sleep insight, motivation, and willingness to use the system, and described a clearer rationale for trying strategies. Our findings suggest that guided reflection made participants more active interpreters of their own sleep. Furthermore, we argue that expert involvement is not a single transfer of knowledge but an ongoing process of making tacit judgment explicit and testable.

0.23

AI-Driven Urge Regulation Assistant (AURA): Designing Preemptive Smart Wearables for Smoking Cessation

cs.HC ·2026-10-07 ·Antoni Timothy, Lam Fu Yuan Kevin, Hou Junyi, Cheong Wei Soon
position
Abstract

Smoking cessation remains a complex self-regulatory challenge, often undermined by cue-triggered cravings arising from habitual, emotional, and social contexts. Existing wearable cessation systems are largely reactive, detecting smoking events only after lapses occur and thus missing the critical window for timely intervention. Guided by cue-reactivity theory and Self-Determination Theory (SDT), we argue that future digital health systems should shift from retrospective feedback toward proactive craving management that supports user autonomy and competence. To inform the design of such systems, we present a formative pilot qualitative study involving semi-structured interviews with 10 smokers and ex-smokers from a multi-ethnic Asian population, a demographic underrepresented in current wearable AI and smoking cessation research. Through an iterative co-design interview process, we examine (1) the barriers and facilitators to smoking cessation and (2) the desired design features of wearable systems targeting cravings before lapses occur. Our findings identify key craving predictors across habitual, emotional, and social dimensions, alongside user-preferred intervention strategies such as social accountability, personalized support, rewards, and just-in-time distraction techniques. Participants also emphasized the importance of trust, privacy, personalization, and non-judgmental interaction styles. These insights directly inform the design of AURA, a novel smartwatch-smartphone ecosystem for proactive craving detection and intervention using commercially available wearable devices.

0.23

Love for Believable AI: Artificial Partiality and Relationship Persistence as an Engineerable Stance

cs.HC cs.CL ·2026-10-07 ·Sebastian Cochinescu
Abstract · 16 pages, 1 figure, 4 tables. Companion framework paper: arXiv:2607.15883 Code and data archived at doi:10.5281/zenodo.21462976

We study a behavioral mechanism for artificial partiality in conversational agents. The paper reports no human-subjects data and makes no claim about attachment, trust, perceived mind, or machine interiority. Partiality has two components: caring, defined as allocation of a finite interaction surplus above a guaranteed per-user service floor, and particularity, defined as a per-user state estimate accumulated from relationship history on fixed inference machinery. We implement both as a persistence layer over one open-weight base model and evaluate them on constructed multi-session relationships with known hidden-state schedules. A four-channel divergence compares known-user and stranger conditions on identical probes with paired generation seeds; the stranger prompt is not length-matched, and one channel (initiative) is an allocator output rather than generated behavior. The full mechanism reproduces the ordinal shape, but not the magnitude, of an analytically specified curve and is the only arm satisfying the protocol-defined joint signature on the calibrated battery. The initial mechanism fails probe-quality equivalence because relationship content reduces topical relevance. A guarded revision satisfies equivalence on the calibrated battery but not on a second battery not used in calibration, where it scores higher than the unmodified baseline. On that second battery, the specificity margin is 0.0198, below the 0.02 protocol threshold. The paired mean right-user--wrong-user estimation-accuracy contrast is 0.375 but is not positive in every run. The mechanism is therefore a candidate for a later perception study, not evidence that users perceive love. The versioned protocol record is not independently time-stamped and is not described as a preregistration. Ethical constraints include a stranger-treatment floor, a bounded surplus, disclosure, and de-intensification.

0.22

A mesh-based neural energy method for the simulation of heterogeneous composites

cs.CE cs.LG ·2026-10-07 ·Pius J. M. Wichmann, Stefan Hildebrand, Sandra Klinge
benchmark
Abstract

Modeling heterogeneous materials remains a challenge for physics-informed neural networks such as the deep energy method (DEM). The DEM and its variants, here collectively referred to as the neural energy method (NEM), offer a differentiable variational framework. However, their conventional collocation-based implementation (C-NEM) often suffers from physically inadmissible displacement oscillations, integration errors, and high computational costs from automatic differentiation. This work introduces the mesh-based neural energy method (M-NEM), extending the NEM through a mesh-based discretization of the displacement field. By interpolating nodal displacements via shape functions, the M-NEM imposes a kinematic constraint that suppresses oscillations. Furthermore, the method replaces automatic differentiation with algebraic shape function derivatives for strain computation and employs high-order Gaussian quadrature for accurate energy integration. On a directly comparable benchmark problem, the M-NEM reduces stress errors by up to three orders of magnitude relative to the C-NEM while being one to two orders of magnitude faster. On two further benchmarks involving extreme stiffness contrasts, only the M-NEM converges. A comparative study of neural architectures reveals that radial basis function neural networks (RBFNNs) yield optimal performance within the M-NEM, resolving sharp gradients at material interfaces with higher accuracy than multi-layer perceptrons (MLPs) with random Fourier feature (RFF) mapping and faster convergence than Kolmogorov-Arnold networks (KANs).

0.22

TutorLoop: Regulating Student Learning Behaviors via Sensor-in-the-Loop Generative Feedback

cs.HC cs.AI ·2026-10-07 ·Songlin Xu, Xinyu Zhang
Abstract

We present TutorLoop, a sensor-in-the-loop system that regulates student learning behaviors by delivering adaptive feedback based on real-time cognitive states. Unlike prior large language model (LLM) tutors that directly depend on scenario-specific content, TutorLoop operates on sensor-derived signals captured via webcams. Moreover, unlike direct cognitive-to-feedback mappings that are short-sighted, the system employs a deep reinforcement learning (DRL) agent to optimize the feedback type across the entire learning process. Finally, another LLM tutor refines feedback into human-like, context-aware messages. We evaluate TutorLoop in a large-scale user study (N=187), where a model trained offline is directly applied to a new learning task without retraining. Results show that TutorLoop provides less frequent yet more effective interventions, improving attention, reducing workload, increasing engagement, and ultimately enhancing learning outcomes. These findings highlight the potential of closed-loop, sensor-driven feedback for scalable human-AI integrated systems to support learning.

0.22

AI-Mediated Self: How HCI Defines and Relates to the Self

cs.HC cs.AI ·2026-10-07 ·Jenny Xiyu Fu, Qian Yang, Malte Jung
Abstract

How might AI alter how we understand and experience the self? This scoping review analyzes 102 papers to examine how the self is defined in the field of human-computer interaction (HCI), how AI-self relationships are conceptualized, and what risks emerge when AI becomes entangled with selfhood. Our synthesis makes three contributions. First, we define AI-mediated self as a conceptual umbrella that connects dispersed work across education, workplace, health, and creative practices. Second, we consolidate six framings of the self with four domains of ethical risk-agency/autonomy, identity/authorship, relational capacity, and meaning-making-into a conceptual map that provides a reusable vocabulary across contexts. Third, we introduce the Inclusion of AI-Self framework, which situates AI-self relationships along a spectrum of proximity. Together, these contributions position selfhood as a central design space in HCI.

0.21

Feeling Wistful: Reflecting on Scholarly Sensibilities with Creative Reading Traces

cs.HC ·2026-10-07 ·Sophia W. Liu, Kate Chier, Shm Garanganao Almeda, Max Kreminski, Bjoern Hartmann
Abstract

Researchers often read before they can articulate what they are looking for. As AI increasingly mediates scholarly search and synthesis, understanding and preserving the idiosyncratic judgments guiding early exploration become important. We call these evolving orientations scholarly sensibilities. To understand curiosity-driven reading, we first examined Wikipedia rabbitholing, a self-directed browsing practice, then designed Wistful, a research probe for open-ended scholarly exploration that captures reading paths as creative reading traces. We studied Wistful with 16 HCI researchers---eight junior and eight senior---in comparison with their usual workflows. Researchers approached the same scholarly landscape differently, with familiarity and personal interests shaping what they pursued and semantic proximity and scholarly links shaping their paths. Their traces made these differences visible and let readers revisit and compare their paths. We position creative reading traces as artifacts for reflection and exchange and as a means of studying how scholarly sensibilities are expressed through reading.

0.20

How assigned AI use before class shapes active student engagement in class

cs.HC cs.AI ·2026-10-07 ·Dan J. Wang, Neelam Modi Jain, Vanessa Burbano, Jorge Guzman, Daniel Keum +3 more
Abstract

AI learning tools are rapidly entering classrooms, but evidence about whether they help students learn is mixed and rests mostly on test scores. Comparatively less research addresses whether the use of AI changes students' live learning behaviors in class. Here, we report the results of a preregistered field experiment with 759 MBA students enrolled in ten sections of a course, in which each student was randomly assigned two of ten class sessions to prepare for with a purpose-built voice-based AI discussion partner. After two uses of the AI discussion partner, students made about 31% more voluntary contributions in each later class session. Students who used the AI discussion partner more also reported greater comfort speaking up and greater perceived learning, but not greater focus or motivation. These findings suggest that repeated practice with a voice-based AI partner can meaningfully increase students' engagement in class discussion, enhancing a critical intermediate learning outcome.

0.20

Comprehension Audits to Mitigate Risks from Automated AI Research

cs.CY cs.AIcs.SE ·2026-10-07 ·Ronald J. Bodkin, Bahrad A. Sokhansanj, Gillian K. Hadfield
code/data
Abstract · 23 pages, 2 figures, 11 tables

AI is already writing a majority of code for frontier AI labs. This creates a safety risk if there is insufficient human oversight. Existing work proposes minimum comprehension thresholds and unaided checks to mitigate this. To our knowledge, however, there is currently no published frontier-AI assurance regime that requires demonstrated evidence that the responsible humans understand what they are building as a precommitted condition for continuing development or usage. We propose comprehension audits, a novel development-process assurance mechanism in which the responsible people explain R&D contributions to auditors to demonstrate understanding. With independent administration and graded reports, they provide a gate: development of a contribution stops based on a failure to demonstrate human understanding until remediated, with escalating consequences for repeated failures. Our analysis of leading open-source AI projects finds increased output of code with reduced human review commentary rates per line of code, with far lower rates for automated fleet accounts. We advocate for labs to conduct them with embedded independent auditors.

0.20

Structured pre-generation elicitation versus single-shot prompting in AI-assisted enterprise decision-making: a randomised online experiment

cs.HC cs.AI ·2026-10-07 ·William Scott-Jackson
Abstract · 12 pages

Generative AI speeds, and mostly improves, professional work, but there is concern that users who delegate both the production and the evaluation of an answer may accept weak output and engage less with the underlying reasoning (cognitive surrender). Interventions proposed so far, such as unassisted practice or slowing adoption, sit outside the working task. We tested a different approach: an interactive metacognitive scaffolding layer (Cognistance, a prototype developed at the Oxford Centre for Impact Research (OCIR) that asks users to clarify context, choose a strategic direction and explain their reasoning before the AI generates a deliverable). Mean composite quality was 32% higher with the scaffold, with the same direction for every rater. Gains were largest for trade-off articulation and strategic coherence and absent for technical specificity. A large part of the aggregate effect reflected rescue of weak prompts: floor-scored (off-task) deliverables fell from 34% to 5%. Among participants whose own prompt already stated the data-localisation problem, the advantage was 21%. Treatment participants reported greater involvement and took about 2.4 minutes longer on average (10.46 minutes). Immediate recall scores were higher, which tentatively suggests better retention, but in this limited experiment, was not robust to sensitivity analyses. Self-ratings of quality did not track rated quality in either condition. Structured elicitation before generation improved the rated quality and task relevance of AI-assisted strategy documents at modest cost in time. Delayed retention, error detection and effects in live organisations are the priorities for the next stage of research.

0.18

Evolve on the Host, Predict on the Edge: Deploying Online Neuroevolutionary Architecture Search for Cross-sectional Stock Return Prediction

cs.CE cs.LGcs.NE ·2026-10-07 ·Jonathan Chang, Zimeng Lyu
Abstract

Accurate forecasting models are usually large, expensive to update online, and fixed in architecture once trained. We apply ONE-NAS, an online neuroevolutionary architecture search that evolves a population of small recurrent networks as each window of data arrives, to daily cross-sectional stock return prediction, and pilot it on a host and endpoint pipeline: the host runs the search and ships each generation's champion genomes over TCP/IP to a Raspberry Pi 4B, which predicts online. On the Pi a single champion predicts a 50-stock window in 24.6~ms and the ensemble of 40 island champions in 556~ms, far inside the daily decision cycle. On four panels of US mid-cap equities over 2022--2024, reading the population as a rank-mean ensemble of island champions returns $+27.5\%$ net of realised transaction costs, against $+11.3$ to $+14.8\%$ for online LSTM, online GRU and monthly-retrained LSTM baselines and $+4.5\%$ for the single best genome used in prior ONE-NAS work.

0.18

Strategic Investment Decision Making for Value Creation in Energy Transition: A Reinforcement Learning Approach

cs.CE cs.LGeess.SY ·2026-10-07 ·Yasaman Cheraghi, Reidar B. Bratvold, Aojie Hong, Ressi B. Muhammad, Sergey Alyaev
benchmark
Abstract

The global challenge of climate change has driven significant steps to reduce CO2 emissions, guided by international agreements like the Paris Agreement of 2015. Acting too slowly could result in future losses and reputational damage, while moving too quickly could jeopardize shareholder value due to the marginal profitability or potential losses due to technology immaturity of many renewable projects. To navigate this complex transition, energy companies must adopt Sequential Decision Making (SDM) strategies to maximize value creation from decision flexibility under uncertainties. To support this, we developed a custom simulation environment to model the dynamic energy landscape up to 2050. Building on this, we designed a multi-criteria SDM framework that explores various decision strategies related to different portfolios for allocating funds across three sectors: oil & gas, renewables, and CO2 reduction. It aims to maximize value during the transition while accounting for uncertainties in productions, energy prices, and costs. This framework has three objectives: maximizing profit, minimizing CO2 social costs, and enhancing competitive advantage in the renewable energy sector. This research evaluates the use of Reinforcement Learning (RL) to identify optimal investment policies within the defined SDM framework. The agent's sequential decisions shape a virtual dynamic environment by influencing key variables such as oil and gas production, renewable energy output, CO2 emissions, and revenues. Through repeated interaction, the RL algorithm explores the state space and learns an optimal policy under uncertainty. We benchmark the RL strategy against a set of manually defined baseline policies and find it consistently outperforms them in adaptability and long-term value creation.

0.17

Before Bringing It Up: When and How AI Companions Should Use Memory

cs.HC ·2026-10-07 ·Zihan Guo, Roxy He, Junwei Quan
Abstract · 17 pages, 1 figure, 3 tables. Zihan Guo and Roxy He contributed equally to this work and share first authorship

Memory can sustain AI companionship, yet even accurate recollection can be inappropriate to use. Two rounds of formative interviews with 14 users (n = 6 exploratory, n = 8 memory-focused) motivate asking what a companion should consider before using past information. Eight themes inform Reconsider, a single-call procedure with five checks and four handling modes, evaluated on 80 scenarios across five models over 400 blinded within-model pairs. Two LLM judges favored Reconsider by net margins of +15 and +23 percentage points, with bootstrap intervals excluding zero for three of five models but not for GPT or Claude. Evaluator analysis linked judge scoring differences to model family, and a preliminary matched-guidance control isolating memory-specific content gave positive margins. We contribute an interview-grounded design framework for memory use and an evaluation that scrutinizes its own evaluators.

0.17

PalmSpace: Towards a Versatile On-Palm Interaction Space through Unified Touch Modeling

cs.HC cs.CV ·2026-10-07 ·Chentao Li, Mingze Gao, Runze Sun, Zhaoguo Wang, Jianjiang Feng +1 more
Abstract · Preprint. Initial version

As smart glasses and lightweight MR devices become increasingly practical, input remains a key challenge. The bare palm is an always-available, tactile, and proprioceptively accessible surface, but it has neither an explicit coordinate system nor embedded touch sensing. Prior on-palm systems typically expose isolated touch events, discrete regions, continuous trajectories, or task-specific gestures, limiting the palm's ability to support precise selection and gesture manipulation through a common input representation. We present PalmSpace, a wrist-worn infrared system that exposes mode-aware, body-referenced absolute input on the bare palm without per-user sensing calibration. At the interaction level, PalmSpace jointly represents contact occurrence, interaction mode, and palm-referenced absolute location; at the model level, it learns these coupled outputs through a shared real-time representation. In leave-one-participant-out evaluation with 17 participants, PalmSpace achieved 6.7 mm mean localization error, 98.9% contact detection accuracy, and 96.7% F1 for four-class interaction-state recognition. User studies further demonstrated absolute pointing and dragging, eyes-free digit input, and representative multi-finger controls including scrolling and pinch-based map manipulation. These results show that a morphologically variable bare palm can function as a transferable, mode-aware interaction surface.

0.17

Conditional Flow Matching for Generation of 3D Multi-variable Instantaneous Urban Microclimate Fields

cs.CE physics.flu-dyncs.LGphysics.ao-ph ·2026-10-07 ·Peng Liu, Shaoxiang Qin, Theodore Potsis, Lili Ji, Dingyang Geng +1 more
Abstract

Rapid and accurate prediction of urban wind and temperature fields is important for urban microclimate design and climate adaptation. Large-eddy simulation (LES) effectively resolves these instantaneous fields, but its application is limited in iterative design of urban microclimate applications due to high computational cost. Existing regressive data-driven models offers quick outputs, but they produce only deterministic point predictions that inherently fail to represent turbulent stochasticity. This paper adopts a novel generative framework of Conditional Flow Matching (CFM) that uses building geometry and mean flow as guidance to generate plausible three-dimensional instantaneous velocity and temperature fields for urban microclimate in seconds. To overcome the GPU memory bottleneck of pixel space 3D generation, the model operates in parallel on overlapping pixel space through a shared-noise initialization that preserves high spatial continuity of flow structure across the entire domain. Against reference LES data, the CFM surrogate can rapidly and accurately restore the first-order statistics with Normalized Root Mean Square Error (NRMSE) of 2.99% for wind and 1.77% for temperature, second-order turbulence metrics with NRMSE of 7.17% for wind and 8.84% for temperature, turbulent kinetic energy with NRMSE of 7%, probability density function and vertical profiles in representative locations. Wind engineering application of local gust prediction demonstrate that the speed and accuracy of CFM, supporting the use of generative AI for making turbulence-aware resilient urban design and climate adaptation more computationally feasible.

0.15

A Scoping Review and Experimental Study on Reinforcement Learning from Human Feedback for Human-Robot Collaboration

cs.HC cs.AIcs.RO ·2026-10-07 ·Alexandra Coroiu, Andrea Vogt, Viktor Werbilo, Andreas Poppele, Johann Christensen +1 more
Abstract

Human-Robot Collaboration (HRC) can facilitate mass customisation in Industry 4.0, with Reinforcement Learning from Human Feedback (RLHF) representing a promising approach for developing safe AI-based robots. Practical challenges remain regarding safety during AI development, human feedback quality, and bidirectional human-robot adaptation. We conducted a scoping review of RLHF in HRC systems, mapping methods that address these challenges. Following PRISMA guidelines, we screened 199 records and included 20 peer-reviewed publications (2020-2025) spanning multiple HRC domains. To our knowledge, this is the first review focused on the bidirectional, closed-loop design of RLHF. Our review found multiple feedback modalities enabling data collection in various feedback formats. Collected data can be integrated at different stages of AI training, resulting in a multi-step development process. Pilot experiments are commonly used to evaluate HRC systems based on both human and robot metrics. To empirically test a key gap identified in the review, we conducted a between-subjects VR experiment comparing system- and user-initiated feedback on robot proxemic behaviour for safe navigation. Using Bayesian models, we analysed the relation between the collected feedback and safety metrics: psychological safety (post-experiment questionnaire) and physical safety (inverse time-to-collision). Results show that user-initiated feedback captures perceived safety better than system-initiated feedback, indicating that feedback timing directly affects feedback quality. Our review and experiment findings show that RLHF relies on appropriate feedback methods to ensure AI safety in HRC, and future RLHF research should prioritise realistic HRC experiments evaluating the effects of feedback collection methods on relevant human and robot metrics.

0.13

Improving social media for democratic discourse

cs.CY cs.HC ·2026-10-07 ·Fan Cheng, Amirhossein Farzmahdi, Pinyuan Feng, Kedar Garzón Gupta, Trenton Jerde +6 more
Abstract · 55 pages. White paper presenting a modular set of mechanisms for designing social media to support democratic discourse

Social media have expanded opportunities for communication and political participation, but today's dominant platforms are optimized primarily for engagement and advertising revenue, contributing to concerns about polarization, misinformation, social isolation, and loss of civility. We explore how social media might instead be deliberately designed to support democratic discourse, collective deliberation, and collective intelligence. Drawing on literature across computer science, psychology, political science, and related fields, we present a modular collection of mechanisms that could be implemented individually or in combination. These include user-controlled and open recommender systems, tools for exposure to diverse perspectives, new forms of cognitive and epistemic feedback, collaborative and AI-assisted fact-checking, privacy and visibility controls, mechanisms for improving civility and evidentiary integrity, and reputation systems that reward high-quality participation. The proposals are intended both as a practical menu of design possibilities and as a starting point for broader interdisciplinary discussion about digital public spaces designed around democratic values rather than engagement alone.

0.11

Human movement reconstruction through latent structural representations during assisted transitions and multidirectional hopping

cs.CE ·2026-10-07 ·Guilherme Manna Cesar, Ligia Yumi Mochida, Chady Ghnatios
Abstract

The quantification of three-dimensional human kinematics is fundamental to neurorehabilitation and musculoskeletal research, yet atypical posture, physical assistance, and rapid movement create conditions of partial observability. Here, we evaluate a latent-structure-informed framework combining convolutional rank-reduction autoencoding with feedforward regression to reconstruct movement from synchronized video. Three case-based demonstrations span therapist-assisted sit-to-stand in a child with cerebral palsy, unassisted sit-to-stand in a typically developing child, and multidirectional single-leg hopping in a young adult. Reconstruction was examined through positional trajectories, projected angular descriptors, and hip-knee coupling. Across test repetitions excluded from model training, mean absolute errors in a predicted marker position during assisted sit-to-stand ranged from 12.02 to 57.11 mm. Sagittal knee and trunk descriptor errors were 5.32° and 5.38° in the typically developing case. Hip-knee cyclograms captured aspects of coupled-motion patterns in both cases, with closer trajectory correspondence in the typically developing case. Sixteen test hops yielded a vertical knee-marker mean absolute error of 15.58 mm. These findings connect latent-structure-informed reconstruction with biomechanically interpretable movement features, providing a foundation for video-derived assessment spanning clinically atypical movement and high-dynamic athletic movement.

0.09

What does it mean to use AI critically? Unpacking critical AI literacy through students' evaluation of AI-generated content

cs.HC ·2026-10-07 ·Hyejeong Lee, Wonjin Yu
Abstract

Although critical AI literacy has emerged as an important educational goal, the construct remains conceptually broad and insufficiently specified for guiding students' day-to-day interactions with AI. This study examines how students critically evaluated AI-generated content. Drawing on an analysis of students' chatbot interactions and written reflections, we first identified seven stages of AI-supported academic task completion and five functional roles assumed by AI. More importantly, we identified eight evaluative lenses that students used to assess AI-generated responses (accuracy, completeness, task alignment/relevance, personalization, practicality, creativity, ethics, and bias). The findings suggest that critical AI literacy extends beyond verifying the factual accuracy of AI-generated information. Rather, critically engaging with AI involves situated and multidimensional judgment about the epistemic quality, contextual fit, practical feasibility, creative value, and ethical implications of AI-generated content. The framework offers both a conceptual contribution to the emerging literature on critical AI literacy and a practical tool for educators seeking to help learners move from passive acceptance of AI toward deliberate, context-sensitive, and responsible human judgment.

0.08

Human-AI Conversational Behaviors Predict Unassisted Task Performance

cs.HC ·2026-10-07 ·Li Siyan, Federico Bianchi, James Zou, Kaitlyn Zhou
Abstract

Assistance from AI tools has supported and improved human performance across domains. However, recent research suggests that these immediate benefits may entail future costs, including diminished performance when AI assistance is no longer available. We study how human-AI interaction behaviors correlate with immediate and future unassisted task performance across two game-based problem-solving user studies ($n=139$ and $n=111$), using a \textit{dialogue act} framework adopted from a tutor-student dialogue taxonomy. In our studies, verbalizing thought processes correlates with higher unassisted outcomes, whereas directly requesting solutions correlates negatively. Similar to these participant-side patterns, assistant explanations of the current problem state are associated with better subsequent unassisted performance, whereas directly providing the next action is associated with worse performance. Qualitative and subtype analyses further show that ostensibly similar reasoning turns can elicit different assistance. Our findings suggest that preserving users' cognitive participation in problem-solving may support performance beyond AI-assisted interaction.

0.08

Beyond Score Accuracy: Examining the Diagnostic Quality of LLM-Generated Structured Assessment in Higher Education

cs.CY ·2026-10-07 ·Xi Zhao, Xinyue Jiao, Zhen Xu
Abstract

As Large Language Models (LLMs) are increasingly adopted for automated grading and feedback in higher education, their structured outputs, including multi-dimensional rubric scores, detailed feedback comments, and improvement suggestions, create an appearance of thorough analytic evaluation. This study examines whether these outputs deliver what they appear to offer. Using the JorGPT dataset of 3,041 student responses to 50 open-ended computer science questions, scored by both human instructors and three commercial LLMs, we identify three systematic discrepancies between the apparent and actual quality of LLM-generated grading and feedback. The sub-dimension scores are highly correlated (r = 0.82-0.99, VIF up to 45), providing redundant rather than independent diagnostic information. The textual feedback rarely detects student misconceptions (5-7% vs. 15-31% for teachers), functioning as a coverage checklist rather than a diagnostic instrument. The feedback tone remains uniformly positive regardless of response quality, lacking the severity modulation observed in human feedback. Additionally, grading accuracy varies significantly by knowledge domain, with procedural topics most reliable. These findings provide empirically grounded guidance on which aspects of LLM-generated grading and feedback can be relied upon and which require continued human oversight.

0.08

AI-Assisted Submissions in Online Research Are Rare and Highly Concentrated but Routinely Approved

cs.HC cs.CY ·2026-10-07 ·Neil K. R. Sehgal, Manuel Tonneau, Dunigan Folk, Lyle Ungar, Sharath Chandra Guntuku
Abstract

Online research platforms underpin much of what science claims about people, on the assumption that a human produced each response. Generative AI threatens that assumption by letting participants delegate responses to a chatbot, yet how often they do so remains unclear because prior estimates rely on self-report or automated detection rather than direct observation. We surveyed 2,500 workers on a high quality online research platform and directly observed AI assistance by linking donated ChatGPT histories to platform submission records of weekly ChatGPT users, covering 712,930 submissions across more than 127,000 studies. Although one in eight surveyed workers reported ever using AI on a study and 68% of observed workers had done so, assistance appeared in only 1% of submissions, with no detectable increase over more than 3 years. Assistance was often temporally localized within tasks and highly concentrated among workers, with 5% accounting for 64% of assisted submissions. Where assistance occurred, workers were virtually always paid, even when study instructions prohibited AI use, with overall approval rates similar to those for unassisted submissions. Half of assisted submissions involved bounded responses, outside the scope of the platform's LLM detector for open-ended responses. Taken together, our results do not support the view that AI assistance currently poses an existential threat to online research, but reveal limited payment consequences for prohibited use and gaps in platform safeguards, leaving platforms poorly prepared should that threat materialize.

0.05

DuoSketch: How Pairs Navigate Challenges in AI-Supported Collaborative Design Ideation

cs.HC ·2026-10-07 ·Weiyan Shi, Darryl Lim, Geraldine Quek, Kenny Tsu Wei Choo
Abstract · work in progress

Generative AI offers new resources for collaborative design ideation, yet how designers work with it as a shared concept develops remains underexplored. We developed DuoSketch, integrating a shared canvas and live transcripts with a separate AI for each member of a pair. An exploratory qualitative study with 12 pairs identified three challenges: unresolved fit between AI proposals and shared design purposes, shared meanings not expressed in AI responses, and agreed changes missing from later outputs. We found that pairs navigated these challenges through bridging work in two directions: AI -> Shared Design (reworking features, developing uses, and assessing fit with shared requirements) and Shared Design -> AI (supplying materials, explaining meanings, and communicating decisions). We discuss how future AI systems can support the joint exploration and articulation of design ideas.

0.04