CS Paper Radar

Daily listing

2026-10-18

409 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 144

A learner-free swarm approximates the explore-exploit equilibrium allocation: A self-organized attractor in place of reinforcement-learning machinery

cs.LG cs.AIcs.SImath-ph+2 ·2026-10-18 ·Galen J. Wilkerson
code/data
Abstract · 26 pages, 13 figures, 5 tables, code and data: https://github.com/galenwilkerson/self-organizing-reinforcement-learning

Reinforcement learning (RL) solves the explore-exploit problem by training a learner on reward; its multi-agent, resource-sharing form has an equilibrium allocation as its optimum -- the ideal-free distribution (IFD), the Nash equilibrium of a congestion game. We show a minimal swarm approximates it with no learner at all. We present a decentralized, learner-free method for this problem: agents explore by heavy-tailed (Levy) flights, recruit neighbours to a discovery through a degree-normalized threshold cascade, and disperse as a patch crowds. The method drives the collective to the shape of the equilibrium allocation (pooled spatial matching slope 1.14, R^2=0.61) with no training, no stored parameters, and no reward in its dynamics; it reaches this as a self-organized attractor and holds it, re-finding it when the resources move -- though it reaches the shape, not the exact equilibrium, and stays measurably exploitable (Nash-gap 1.46). Coupling recruitment to movement causes the allocation: removing it collapses the matching (slope 0.006). Against Proximal Policy Optimization (PPO) learners (feedforward, centralized-critic, graph message-passing, and recurrent) trained in the same sensing-limited environment, the method at zero training is the only arm left unbiased w.r.t. the IFD gradient (slope ~1, CI containing 1, vs learners 0.69-0.97, all under-concentrating) and ties the best learner's harvest where resource is clustered, while the learner collects more where resource is scarce (0.33 vs 0.22) and holds a tighter per-capita intake. Learning buys precision and transfer across reward structures, not a better allocation shape. A companion paper develops the mechanism.

0.62

Formal Raising: Embedding Verifiable Ethical Properties into the Development of AI

cs.AI ·2026-10-18 ·Rolf Drechsler, Christina Plump
Abstract · 11 pages, 6 figures,

Alan Turing proposed constructing a ``child machine'' and educating it rather than directly programming an adult-like intelligent machine. Recent work on model raising revives this developmental view for Large Language Models by arguing that moral values should accompany capability acquisition from the beginning rather than be added only through post-hoc alignment. This paper proposes Formal Raising (FR): selected moral values are represented as explicit, machine-checkable properties that participate in development, evaluation, and action selection. We present a C-based reasoning framework in which candidate actions are evaluated against prioritized ethical properties. Conflicts are resolved w.r.t. a given order, strict rules are non-defeasible, and underdetermination is reported explicitly. Exceptions are not encoded as special cases but emerge from the comparison of alternatives. We provide examples that cover conflicts involving life, truthfulness, harm, privacy, autonomy, and strict rules. We outline a property layer inspired by a Property Specification Language (PSL) for state and temporal requirements and discuss its integration into a raising process.

0.54

Narrow Decision Models: A System-1 Architecture for Calibrated, Typed Decisions

cs.AI cs.CLcs.LG ·2026-10-18 ·Carlos André Antunes, Aldry Albuquerque, Gabriel de Melo Schmith, Natã Batista, Pedro Oziel Escobar da Silva +5 more
hot direction
Abstract · 27 pages, 6 figures, 4 tables. A first training on real documents is summarized; full results on real data in production are left to a compa…

System One models, introduced with TypeSafe's Jev, answer typed questions about a state with calibrated probabilities instead of generated text: choosing among options described in natural language (Choice), placing the state on an ordered scale (Score), or judging a yes/no statement (Noul). Jev is general-purpose: one set of weights, served as an API and adapted to a domain only through the request. We study the other end of the design space. A Narrow Decision Model (NDM) offers the same interface but is trained for the decisions of one organization, is small enough to run on its own infrastructure, and has an open architecture whose properties can be stated and tested. The state can be any text or structured record about which decisions are made; our running case is the routing of court communications. The architecture serializes a request as a prefix tree (the state, then each question, then each option) and runs a bidirectional encoder under a block attention mask with logical positions, so that each answer is, by construction, invariant to the order of its options and independent of the other questions asked. Routing through an organization chart is a composition of Choice decisions with an explicit "none of these" option. We describe training with proper scoring rules through a curriculum, calibration and abstention, decisions composed by rubrics, and a protocol that accepts or rejects the mechanism before held-out data is opened. A minimal validation on public and fictitious data shows that the invariances hold to rounding in a trained model, that a small instance routes on organization charts it has never seen well above lexical retrieval, and that, at equal compute, the block architecture reaches a lower loss than a pairwise cross-encoder. A first instance trained on real legal documents agrees with the LLM classifier in production on 91% of held-out documents.

0.53

ClinSDT: LLM-Encoded Clinical Semantic Guidance for Sepsis Treatment via Decision Transformer

cs.LG cs.AI ·2026-10-18 ·Yanyan Wu, Guanyu Zhu, Changbin Zhao, Jin Zhang, Dexu Yu +5 more
top venuecode/data
Abstract · 12 pages, 6 figures. Accepted at CIKM 2026 (Full Research Track, Oral Presentation)

Offline sepsis treatment aims to learn treatment policies from retrospective ICU trajectories. Decision Transformer-based methods provide a natural sequence modeling framework for this task, but they mainly rely on raw physiological states and historical clinician actions. Such raw trajectory tokens may not explicitly capture clinical semantics that clinicians use in practice, such as patient severity, recent treatment response, and current risk priorities. To address this limitation, we propose ClinSDT, an LLM-assisted Decision Transformer framework for offline sepsis treatment. ClinSDT uses a frozen LLM as a semantic bridge rather than a treatment decision maker. It constructs three clinical semantic views from ICU trajectories, namely Profile, Context, and Advisory, and inserts their encoded representations as independent tokens into a Decision Transformer policy. To support policy refinement and rollout, ClinSDT further trains a Treatment-Response Forecaster to estimate the next physiological state, SAPS-II change, and predictive uncertainty from state-action inputs. In the second training stage, the forecaster screens local counterfactual treatment candidates, and only low-uncertainty candidates with lower predicted SAPS-II increase are distilled into the policy after semantic regeneration. During inference, ClinSDT performs closed-loop treatment rollout by refreshing semantic inputs from the currently available patient state. Experiments on MIMIC-III and MIMIC-IV show that ClinSDT achieves consistent relative improvements over representative imitation learning, offline reinforcement learning, and Decision Transformer baselines under the same rollout protocol. Code and data are available at https://github.com/KAG778/ClinSDT.

0.52

BrandEval: A Two-Track Multi-Agent Benchmark for Risk-Sensitive LLM Crisis Communication

cs.AI ·2026-10-18 ·Haotian Gan, Weidong Tang, Hanbin Sun, Yi Luo, Xintian Li +3 more
top venuebenchmarkcode/data
Abstract · Accepted to EMNLP2026 Findings

Evaluating LLMs for crisis communication requires more than judging isolated responses, because public feedback and competitor actions can reshape trust over time. Existing benchmarks remain largely static and single-turn, offering limited evidence about long-horizon risk-sensitive communication. We introduce \textbf{BrandEval}, a two-track benchmark for product-crisis communication, focusing on public brand responses to product-related safety, quality, service, and trust incidents. BrandEval-Static evaluates stakeholder-aware strategic analysis, while BrandEval-Dynamic stress-tests response policies in \textbf{BrandPolis}, a controlled multi-agent sandbox with consumer feedback, competitor responses, and trust-risk trajectories. We also construct \textbf{BrandSRD}, a Chinese crisis-response dataset, and instantiate a reference Strategic Agent using \textbf{Strategic Rationale (SR)}. Across proprietary and open-source models, we find that static strategic quality is informative but incomplete: it aligns with some end-state outcomes but does not reliably predict volatility and delayed recovery; dynamic stress tests further expose tail-risk failures under manipulative or weakly grounded policies. The code and dataset will be available at https://github.com/htgan-ai/BrandEval.

0.50

SCOPE: Source Credibility and Outlet Profiling Engine

cs.MA cs.AIcs.CL ·2026-10-18 ·Ilyas Nagiyev, Miras Baisbay, Daniil Orel, Preslav Nakov
top venuebenchmarkcode/data
Abstract

We present SCOPE, a multi-agent system for profiling news media outlets with respect to bias, factuality, and credibility. Given an outlet URL, SCOPE coordinates specialized agents for article acquisition, content analysis, external evidence search, scoring, and report generation. The system follows an MBFC-style methodology while making intermediate evidence and agent outputs inspectable through a structured demonstration interface. On a benchmark of 200 media outlets, SCOPE outperforms LLM-only, article-only, search-only, and media-background-check baselines on most metrics, achieving the lowest bias and factuality MAE, perfect failed-fact-check detection, and the lowest factual error rate. These results suggest that decomposed, evidence-grounded agent workflows can support more transparent and reliable media profiling. Our code is available at https://github.com/mbaisbay/SCOPE, and a live demonstration at https://scope-demo.streamlit.app/.

0.49

Paloma: Phase-Conditioned Residual Modulation for Time Series Forecasting

cs.LG ·2026-10-18 ·Jingru Fei, Wei Fan, Kun Yi, Qi Zhang, Zhendong Niu
top venuecode/data
Abstract · Accepted by NeurIPS 2026 spotlight

Explicitly modeling periodicity has proven to be an efficient approach for time series forecasting, where the residual, obtained after removing the periodic component from the time series, is typically treated as unstructured noise. However, in this work, we show that such residuals still exhibit structured, \textit{phase-dependent} deviations, as supported by both empirical observations and theoretical analysis. This suggests that residuals should not be treated merely as unstructured noise, but instead as \textit{phase-conditioned signals} whose variations arise in both direction and magnitude. Motivated by this perspective and inspired by phase and amplitude modulation in classical signal processing, we propose \textit{Paloma}, a simple yet effective framework for modeling residual deviations. Specifically, Paloma comprises two complementary modules: \textit{a Phase Rotation module}, which performs phase-conditioned rotations to capture directional changes in the feature space, and \textit{an Amplitude Scaling module}, which applies phase-conditioned affine transformations to model magnitude variations. We further abstract the two modules into a general-purpose plugin, termed the \textit{Paloma technique}, which can be readily integrated into existing forecasting architectures to modulate input sequences in a phase-aware manner, enabling backbone models to capture phase-dependent variations in both direction and magnitude. Extensive experiments on twelve datasets demonstrate that Paloma achieves state-of-the-art and robust forecasting performance. Additional results show that the Paloma technique consistently improves a wide range of baseline models, validating its effectiveness as a general-purpose component. Our Code is available at \url{https://github.com/TSTS13/Paloma}.

0.48

P(Yes): 15$\times$ Improvements via Test-Time Scaling of Decision Models

cs.AI ·2026-10-18 ·Lev Stambler
code/datahot direction
Abstract · 22 pages, 5 figures. Interactive version: https://research.tearlabs.ai/pyes-search/ . Code and data: https://github.com/tear-labs/pyes-searc…

Decision models answer typed questions, such as multiple-choice or yes/no questions, with a probability for each option. We add test-time search to frozen decision models, with no training: we show the model a complete proposed answer, read off its probability of saying "Yes" to the proposal's correctness, and search for answers that it rates higher. The search is done either by editing one answer at a time or by joint optimization, which changes several answers at once. Search raises accuracy for every model and test set we tried, and the gains are largest for exact correctness on logic puzzles, where every answer in a task must be right. On 128 such puzzles, it takes Qwen3.8-27B from 2 to 30 exactly correct puzzles ($15\times$) and its per-blank accuracy from 27.8% to 48.5%. On a shared set of 47 puzzles, searching over each model's own Yes/No score raises the per-blank accuracy of StartLux Instant, Perplexity Decider, Clef, and JEV, which we reach only through its public API, by 4.0 to 13.0 percentage points, and raises the number of puzzles they get exactly correct, out of 47 and averaged over three search seeds, from 10 to 35, 9 to 26, 2 to 11, and 9 to 31.7 respectively.

0.45

SteadyThought: Mitigating LLM Under-Thinking via Thought-Level Preference Optimization

cs.AI ·2026-10-18 ·Zhenyue Peng, Lin Yang, Jie Wang, Xiaoqi Ni, Hanzhu Chen +5 more
top venuecode/data
Abstract · Accepted at NeurIPS 2026

Flexible switching between reasoning trajectories (i.e., thought switching) has significantly enhanced the reasoning capabilities of Large Reasoning Models (LRMs). However, existing models often switch excessively yet fail to sustain promising reasoning thoughts, a phenomenon termed "under-thinking". While recent efforts suppress switching to mitigate this, such over-correction may discard valuable trajectories. To address this challenge, we propose Steady Thought (ST), a novel thought-level preference optimization framework. ST formalizes under-thinking as a preference issue at switching points, forcing single-path continuations from identified points to yield high-quality reasoning. Then, ST performs thought-level preference optimization by treating the newly generated response as preferred and the original one as dis-preferred. Experiments across multiple models and datasets show that ST effectively reduces token consumption, achieving reductions of up to 63.8%, while maintaining strong reasoning accuracy. Further analysis suggests that ST helps models acquire a generalizable ability to reduce redundant switching across domains and languages. Code is available at https://github.com/MIRALab-USTC/SteadyThought.

0.44

KPI-Aware Multimodal Anomaly Detection for 5G and Open RAN Telemetry

cs.LG ·2026-10-18 ·Chiman Salavati, Liang Wu, Kelly Wan, Mayank Darbari, Liangjie Hong
top venuebenchmark
Abstract · Accepted at the IEEE ICDM 2026 Applied Research Track

Reliable anomaly detection is a deployment bottleneck for 5G/Open RAN observability and emerging self-optimizing 6G systems, where hundreds of heterogeneous KPIs, severe class imbalance, and strict alert budgets make purely numerical scores brittle. We present KAC, a deployment-oriented KPI-aware multimodal anomaly-detection framework that uses KPI-level residual-text alignment to turn compact operator-style summaries into a learnable feedback signal over full-KPI temporal evidence. This gives the detector broad numerical coverage like modern time-series models, while retaining the sparse KPI-specific reasoning that makes operator feedback sustainable in a future operator-facing workflow. Across public telecom telemetry, enterprise-scale Open RAN observability data, and a Nokia benchmark built from real 5G RAN production packet traces with realism-hardened injected anomalies, KAC achieves strong threshold-based detection performance while remaining competitive on threshold-free ranking metrics. On TelecomTS, at a fixed 2.5 percent false-positive-rate operating point, KAC achieves the highest recall among the evaluated residual-based methods while maintaining comparable precision; across the evaluated benchmarks, it remains robust under substantial class imbalance. KAC is integrated in pre-production shadow mode with Nokia's GKE-based KPI construction pipeline, where it continuously scores telemetry windows alongside an internal packet-level anomaly detector for operational evaluation.

0.43

Revealing what benchmark scores hide: Jointly measuring the breadth and reliability of LLM capabilities using accessibility profiles

cs.AI cs.CLcs.LG ·2026-10-18 ·Kazem Faghih, Yize Cheng, Shoumik Saha, Mobina Pournemat, Armin Gerami +1 more
benchmarkpositioncode/data
Abstract · 26 pages total, 10 pages main text, 18 figures. Code and data: https://github.com/kazemf78/access-eval

Language model benchmarks assign each question a single outcome from one fixed wording, even though the same underlying request can be expressed in many equivalent ways. We argue that the quantity of interest should therefore not be correctness on one wording, but the accessibility of an inquiry, which we define as the probability that a model succeeds across meaning-preserving formulations. The distribution of accessibility across inquiries reveals a structure that accuracy and existing best-, worst-, or average-case prompt scores collapse: how broadly a model's capabilities extend across inquiries and how dependably those capabilities can be elicited. Two models can have the same average performance while one succeeds across more inquiries unreliably and the other succeeds across fewer inquiries but more consistently, leading to opposite preferences depending on whether the task allows multiple retries until one success or when repeated consistent success is required. Across six benchmarks and more than twenty models, this tradeoff is common: for about half of the model pairs with nearly equal mean accessibility, which model looks better depends on whether one success across five formulations is enough or all five must succeed, quantities we call Access@5 and Reliable@5, respectively. For some pairs the two measures diverge significantly, by up to 10 points in opposite directions. Accessibility also reveals that scaling, reasoning, and prompt optimization can redistribute capability across inquiries in ways aggregate accuracy obscures, sometimes broadening the set of inquiries a model can succeed on while making success on them less repeatable. Accessibility turns formulation sensitivity from a nuisance into a measurable property of capability, distinguishing what a model can sometimes succeed on from what users can actually depend on.

0.43

Shallow Probes Are Sufficient for Near-Oracle Personalization Depth Allocation in Federated Learning

cs.LG ·2026-10-18 ·Beyazit Bestami Yuksel, Emrah Dikbiyik, Ayse Yilmazer Metin
publishedhot direction
Abstract · 6 pages, 9 figures. Accepted and presented at SINCONF 2026 (18th International Conference on Security of Information and Networks), Beykent …

Rather than proposing a new federated optimizer, we study whether a small client-specific measurement budget can support cost-aware personalization-depth decisions under a utility that trades Macro-F1 against deployment cost. Exhaustive evaluation of six depths defines an offline oracle that is not deployable. We audit the Probe Sufficiency Hypothesis: three shallow probes (P0--P2) may contain most decision-relevant information for approaching that oracle. FedPBA is a secondary learned instantiation for deep-action extrapolation. On CIFAR-10/100 and patient-disjoint PTB-XL (45 sessions; 900 clients; 2700 evaluations; GroupKFold), training-free greedy probing achieves regret 0.0071 and 98.3 percent oracle recovery, beating metadata-only policies and FedPBA (0.0110 / 97.3 percent). Aggregate shallow-best versus oracle correlation is 0.9818 but falls to 0.578 on deep-oracle rows; PTB-XL deep-oracle cases remain a high-regret boundary without a validated escalation gate.

0.43

DAGLoop: Aligning Method Design with Experimental Feedback in Automated AI Research

cs.AI ·2026-10-18 ·Dengyuxing Wang, Yonggang Zhang, Qianshu Cai, Huajiang Zheng, Jun Song +2 more
code/data
Abstract · 26 pages, 4 figures, 6 tables. Submitted to ICLR 2027. Code: https://github.com/deng34567/DAGLoop

Automated AI research agents aim to improve machine learning methods by iteratively proposing, implementing, and evaluating candidate methods. Existing work uses the outcome of each evaluation to determine how the next candidate should be revised for performance improvement. However, a candidate often combines multiple design decisions, and its overall outcome alone does not reveal which decisions produced their expected effects, whether their required conditions were satisfied, or which decisions should be retained or revised. Inspired by ablation studies in the machine learning community, we propose DAGLoop to test the expected effect of a design decision by comparing implementations with and without that decision while preserving the other decisions it requires. Specifically, DAGLoop translates each design decision's rationale into a testable expectation and pairs it with a targeted experiment, forming an experimental specification for assessing the expected effect under its required conditions. Using these specifications, we formulate ablation planning as the construction of a directed acyclic graph (DAG) from the candidate method's design: nodes associate decisions with tests, and edges encode dependencies. The DAG specifies which decisions each experiment must retain and supports joint tests when an expected effect requires several decisions to act together. Outcomes are recorded with the tested decisions, expectations, and conditions to form structured research memory. This memory guides the agent in retaining experimentally supported decisions, revising unsupported ones, and planning further experiments when the evidence is unresolved. Experiments on dozens of research tasks demonstrate substantial performance improvements achieved by DAGLoop across diverse executable research settings. Code and evaluation scripts are available at https://github.com/deng34567/DAGLoop.

0.42

ControlJEPA: Principled Trajectory Regularization via Lyapunov Tube Loss

cs.LG cs.AI ·2026-10-18 ·Thai-Khanh Nguyen, Nhi Y. Pham, Uyen N. B. Vo, Thieu N. Vo, Cuong Pham +1 more
publishedcode/data
Abstract · NeurIPS 2026

Next-token prediction leaves hidden-state trajectories largely unconstrained, while existing geometric regularizers primarily control the direction rather than the magnitude of trajectory deviations. We introduce \emph{ControlJEPA}, a trajectory regularization framework based on a discrete-time Lyapunov contraction condition from practical Input-to-State Stability theory. Its Lyapunov Tube Loss penalizes violations of this condition on a normalized, scale-invariant measure of transversal deviation. When the contraction condition holds at every step, the hidden-state trajectory remains within an explicit tube around the reference secant. In the partial-ISS regime, we bound mean transversal energy, temporal variance, and average tube width using the Lyapunov loss and contraction parameters, with horizon-independent variance bounds under uniformly bounded violations. ControlJEPA requires no architectural modifications. Across six benchmarks and six model families, it consistently improves accuracy and data efficiency over standard fine-tuning and prior trajectory regularizers. The resulting smoother transversal profile also persists empirically during inference, when the regularizer is inactive. The code is publicly available at: \url{https://github.com/t-khanusa/ControlJEPA}

0.41

Explore or Confirm? LLM Agents Misallocate Measurements Under Noise

cs.AI cs.MA ·2026-10-18 ·Abdul Karim, Matt Keon
benchmarkcode/data
Abstract · 38 pages, 12 figures. Code, data and environment: https://github.com/Xstendyl/wellocate

As model reasoning gets cheaper, the physical experiment becomes the slow and costly step in discovery. Each experiment returns a noisy reading: a high value can mean a good candidate or a lucky draw. With a fixed budget, an AI agent must keep choosing between exploring a new candidate and confirming a promising one by re-testing it. We measure that choice in Wellocate, a simulated assay campaign with known true values. It is built on a generated landscape and on two real protein datasets, GB1 and PhoQ, scrambled so no model can answer from memory. We test five language models as minimal agents, with no tools and no memory beyond the history in each prompt. They misallocate in both directions. Three over-confirm, spending most of the budget re-testing. On the generated landscape one under-confirms, trusting single readings, and there every model ends with a worse final pick than a basic scripted strategy. Each failure has its own cause. The over-confirmers run short of new candidates worth trying: a list of untried candidates cuts their re-testing and, on real protein data, makes two of them pick much better. A running summary of their readings or a description of the landscape changes it by at most five percentage points. For the under-confirmer, re-testing its top candidates in a final round, replayed on its own campaigns, recovers only about a quarter of what it loses by picking the wrong candidate. This suggests confirmation has to run through the whole campaign. We release Wellocate as a benchmark and reinforcement-learning environment with frozen splits and scripted baselines. With it, anyone can measure how an agent splits its budget before trusting it with real experiments, and train agents to split it better. Code, data and the environment are available at https://github.com/Xstendyl/wellocate.

0.41

ViCoR: Estimating Visual Necessity via Counterfactual Residuals for Multimodal Medical Data Selection

cs.LG cs.AI ·2026-10-18 ·Haotian Gan, Yudong Li, Xintian Li, Jiarui Han, Weidong Tang
top venuebenchmarkcode/data
Abstract · Accepted to NeurIPS2026

Medical vision-language models often exhibit hidden stratification, where high overall performance masks failures on subtle visual findings and overlapping clinical conditions. We trace this failure mode to modality imbalance during fine-tuning, where strong linguistic priors can suppress visually grounded learning and induce plausible but weakly grounded rationales. Existing data selection methods usually score image-text pairs as holistic units, without separating visually necessary samples from shortcut-solvable ones. To address this problem, we propose ViCoR, a counterfactual-residual-based data selection method for Med-VLM fine-tuning. ViCoR estimates visual necessity through image-dependent optimization residuals, calibrates the resulting scores with medical supervision compatibility, and constructs a diversity-aware fine-tuning subset. Across two backbones, four external generalization benchmarks, and one multi-disease robustness benchmark, ViCoR-selected 20\% subsets achieve the strongest average performance compared with full-data fine-tuning and selection baselines. Gradient-level analysis and perturbation audits further show that ViCoR enriches image-dependent training signals and strengthens image-level and evidence-region dependence. The code will be available at https://github.com/htgan-ai/ViCoR.

0.40

Block Recursive Transformers for Structured Parameter Sharing

cs.LG ·2026-10-18 ·Josep Marques, Hyunyoung A. Cho, Youngeun Kim, Youjia Zhang, Aecheon Jung +2 more
top venueposition
Abstract · 13 pages, 9 figures. Published at EMNLP Findings 2026. Typo correction from camera-ready manuscript. Project website: https://aim-skku.githu…

Recursive Transformers depend not only on how many parameters are shared, but also on how shared parameters are organized across depth. Existing static schemes adopt two extreme reuse topologies: layer-wise repetition, where each layer is reused independently, and global cyclic reuse, where the entire stack is repeatedly applied as a single shared function. We argue that sharing granularity is an important factor in recursive parameter sharing. This study proposes Block Recursive Transformers (BRT), which partition layers into contiguous blocks and recursively reuse each block before moving to the next. This intermediate topology enables local refinement within blocks while preserving depth-wise transitions, without adding routing modules, auxiliary states, or extra training objectives. Across decoder-only models up to 1.7B parameters, with additional encoder-decoder and encoder-only studies, BRT improves validation negative log-likelihood and downstream accuracy over static recursive sharing baselines. Ablations and representation analyses further show that intermediate block granularity better balances local refinement and depth-wise differentiation. Project website: https://aim-skku.github.io/BRT/

0.40

HCFM: Helmholtz Component Flow Matching for Time Series Anomaly Detection

cs.LG cs.AIcs.DB ·2026-10-18 ·Shyam Sundar Saravanan, Tie Luo, Sanjay Madria
top venuecode/data
Abstract · To appear as a conference paper in ICDE 2027

Flow Matching (FM) is an efficient simulation-free approach for training flow-based generative models, making it a promising foundation for scalable time-series anomaly detection (TSAD). However, standard FM learns a monolithic vector field in which nominal transport and local density-changing behavior are entangled. This creates a central challenge for TSAD; local expansion and contraction carry crucial anomaly-sensitive information, yet isolating this signal requires expensive Jacobian-trace estimation over the entire field. We propose \emph{Helmholtz Component Flow Matching} (HCFM), an unsupervised TSAD framework that parameterizes the vector field prediction network with separate transport, compression, and residual heads. The transport head is constrained to be divergence-free, so divergence estimation is restricted to the compression head, while a lightweight residual head absorbs the remaining mismatch. This structured parameterization improves scalability and detection performance by separating volume-preserving nominal transport from density-changing compression, yielding cleaner anomaly signals from distinct deviations of nominal behavior. Experiments on large-scale univariate and multivariate TSAD benchmarks show that HCFM consistently improves detection accuracy while maintaining efficient inference compared with 16 baselines. The code is available at https://github.com/TrustAIoT/HCFM__time-series-anomaly/

0.39

FracTS: Hierarchical and Autoregressive Time Series Generation

cs.LG ·2026-10-18 ·Jiayu Li, Umair Afzal, Zilong Zhao, Milad Abdollahzadeh, Uzair Javaid +1 more
code/data
Abstract

Time series are ubiquitous across different domains, and the growing use of synthetic data for privacy-preserving sharing, data augmentation, and stress testing makes time series generation increasingly important. Early generative approaches are predominantly autoregressive, which can suffer from low sample diversity, error accumulation, and difficulty modeling long-range dependencies. In this paper, we introduce FracTS, a hierarchical, coarse-to-fine autoregressive model for multivariate time series generation. Inspired by fractal generative modeling developed for images, FracTS adapts the paradigm from 2D to 1D sequences by learning a multi-scale representation that captures global trajectory structure at coarse scales and local dynamics at fine scales while following the autoregressive nature of time series. In contrast to unstructured images, time series are structured, allowing easy automated or designated temporal feature processing optimized for the model, and global (static) feature extraction describing the instances that generate the corresponding time series in the collection. We incorporate them in the generation process to further improve the generation quality. The design naturally handles multivariate dependencies and improves long-horizon temporal coherence without requiring fixed-length inputs. Experiments on five datasets (two conditional, three unconditional) conducted against seven baselines show that FracTS consistently outperforms baselines across marginal distribution and predictive ML utility, and correlation preservation in the unconditional setting. In particular, the predictive ML utility is averagely improved by 100.0% and 75.0-81.0% for unconditional and conditional time series datasets compared to the best baseline (relative error computed on dataset-averaged raw values). Code is available at: https://github.com/BetterdataLabs/fracts.

0.39

When Form Changes but Logic Doesn't: Building Logic-invariant LLMs through Logical Structure

cs.AI ·2026-10-18 ·Xuyuan Liu, Xinshuai Dong, Elynn Chen, Yujun Yan
top venue
Abstract · Accepted at NeurIPS 2026, 24 Pages

Large language models (LLMs) have achieved strong performance across diverse reasoning tasks, yet it remains unclear whether this performance reflects an understanding of underlying logical structure or a reliance on shallow semantic cues. A key indicator of such understanding is whether models treat logically equivalent inputs consistently, producing the same predictions despite differences in expression. Through systematic evaluation on curated data, we show that even strong LLMs often fail this test: they do not consistently preserve predictions across logically equivalent inputs and generalize poorly to unseen logical forms. To address this limitation, we propose LoGIcal STructure-guided reasoning (LoGIST), a framework that guides LLM reasoning with explicit logical structure. LoGIST maps each input instance to a decision diagram, a directed acyclic graph that captures the underlying logical structure, and encodes this graph into a logical embedding that conditions the LLM's reasoning process. Across multiple training settings, LoGIST reduces inconsistency under logical equivalence and improves generalization to unseen logical forms, suggesting that logical invariance is a useful target for building more robust and generalizable logical reasoning in LLMs.

0.38

Auditing Tool Traces in Agentic LLMs: Source Localization and Counterfactual Dependence Testing

cs.AI cs.CL ·2026-10-18 ·Junhao Liu, Tianxiao Li, Yongxian Wu, Ziheng Duan, Ren Li +2 more
publishedbenchmarkhot direction
Abstract · 12 pages. Accepted at the NeurIPS 2026 Workshop on Agentic AI for Biological Discovery (AgenticLS)

Agentic LLMs increasingly support scientific prediction by grounding decisions in external evidence. In biology, they promise accurate, interpretable predictions for important problems by integrating complex experimental and computational evidence. However, serialized tool results can contain hidden label shortcuts, such as dataset identifiers or template artifacts, that create apparent success without biological reasoning. To address this risk, we introduce a two-stage audit. Stage 1, Per-Tool TF-IDF Decomposition (PT2D), localizes label-predictive patterns to individual tool results while avoiding signal dilution in the concatenated tool trace. Stage 2, Token-Class Redaction, counterfactually tests whether the frozen LLM depends on the localized shortcut. We apply this audit to TCR-pMHC binding, where tool augmentation substantially outperforms a no-evidence baseline. At face value, this suggests that the LLM has learned to use scientific evidence. The audit overturns this interpretation: PT2D localizes the strongest signal to identifiers in one tool result, and redacting them reduces performance to chance. These findings show how hidden label shortcuts can masquerade as scientific reasoning or discovery, motivating source-level audits before benchmark gains are trusted.

0.38

Simulating Infant Senses with Age-Tuned Networks

cs.LG cs.HCq-bio.NC ·2026-10-18 ·Somesh Mohapatra
code/dataunusual
Abstract · 21 pages, 7 figures, 2 tables. Code and website: https://github.com/pikulsomesh/developmental-world-models and https://pikulsomesh.github.io…

What a newborn sees, hears and feels, and how this changes month by month, has been measured sense by sense, with different methods and in different units. We bring these measurements together for the first two years of life and ask whether small neural networks tuned to a given age can turn them into a quick, inspectable simulation of sensory perception. Each sense's immaturity is written as a fixed, age-indexed filter on the input, and a small prediction network learns behind it over the first year. The filters alone reproduce the order in which the senses mature, by construction. The networks add what filters cannot: in vision they extract far less than their input allows at early stages (at the newborn stage a toy's position cannot be read out, $R^2 = 0.00$, against 0.57 for a network trained directly on the same input), and a critical period and perceptual narrowing emerge without being built in; in hearing they track what their input allows. The same filters drive the Baby Senses Lab, a free website that shows a user's own images and sounds as a baby of a chosen age might sense them. It is a research and teaching tool, not a diagnostic test.

0.37

MLFont: Machine Learning-Driven Font Subsetting and Range-Compact Unicode-Range Slicing for Low-Latency Web Typography

cs.LG ·2026-10-18 ·Quan Wang, Yiran Mao
code/data
Abstract

CSS3 @font-face rules with unicode-range descriptors partition web fonts into downloadable slices. While Latin fonts need ~20 KB in WOFF2, Pan-European (LGC) and East Asian (CJK) fonts contain 2,297-30,759 codepoints (158 KB-5.9 MB in WOFF2), where heuristic partitioning causes unused-glyph downloads, high HTTP RTT, CSS bloat, or missing-glyph fallbacks. We present MLFont, a machine learning and optimization framework that learns font slicing strategies from corpora via three stages: (1) a Zipfian High-Frequency Tier (most) isolating top-frequency codepoints; (2) a Latent Co-Occurrence Topic Tier using Online Variational Bayes LDA (Online LDA) or a Deep Neural Network Topic Model (DNNTM) with Shannon Entropy Difference Loss and Sinkhorn-Knopp normalization; and (3) a Memetic Range-Compact Genetic Algorithm (GA) with Potts-Viterbi dynamic programming and island bridging that minimizes topic discrepancy and CSS fragmentation. Across five Google Noto fonts (LGC, KR, JP, TC, SC), MLFont cuts 3G per-glyph latency by 56.7%-75.3% over single-file delivery and 46.1%-66.3% over equal-bin baselines on CJK with 100% coverage, reduces CSS intervals by 17.9%-41.7%, and achieves 1.56x-6.09x cold-cache speedups (65.4%-84.9% CJK and 44.7% LGC bandwidth savings) across 20 websites in simulator and Headless Chrome benchmarks (45/45 exact slice-set parity). Code: https://github.com/wq2012/mlfont Models: https://huggingface.co/collections/wq2012/mlfont

0.37

UtoMe: Observation-Uncertainty-Guided Token Merging for Weather Foundation Models

cs.LG cs.AIcs.CV ·2026-10-18 ·Yunbei Zhang, Janet Wang, Xi Xiao, Jihun Hamm, Xiao Wang
top venuecode/data
Abstract · Accepted to NeurIPS 2026. Code and project page: https://ornl-utome.github.io/

Weather transformers process dense reanalysis fields whose reliability varies across variables, locations, and weather regimes. Token merging can reduce computation, but feature similarity alone does not account for this heterogeneous input uncertainty. We introduce UtoMe, an observation-uncertainty-guided token merging framework that combines dynamics saliency with ensemble data assimilation spread as a relative input-reliability proxy. The resulting importance score protects selected tokens as anchors and calibrates bipartite matching between the remaining tokens. Three variants use external uncertainty at both training and inference, during training only, or neither stage. In ERA5 72-hour forecasting at two spatial resolutions, UtoMe improves Overall error over train-time ToMe at matched token budgets. At $1.40625^\circ$ and 50% token reduction, UtoMe-E lowers Overall error from 67.32 to 53.00 at the same 0.55x encoder FLOPs. Small reductions remain close to the full-token baseline; larger reductions incur measurable accuracy losses. Component ablations and spatial uncertainty controls support combining dynamics with aligned uncertainty when allocating a weather transformer's token budget.

0.35

StopLoss: Teaching Agents to Predict Failure and Stop Early

cs.AI ·2026-10-18 ·Jiarong Liang, Xueguang Ma, Dongfu Jiang, Yuxuan Zhang, Ping Nie +1 more
Abstract

Coding agents keep working on runs that will fail verification, and asking a frontier model to watch for this is expensive in its own right. We introduce StopLoss, a small external monitor that learns failure recognition from completed trajectories and their verifier outcomes. The name is inspired by the stop-loss order in finance: StopLoss ends a run early when the run is likely to fail, which saves resources. Final outcomes label earlier local windows, so a fine-tuned language model can score the most recent tool steps in a single forward pass, and a threshold calibrated on successful runs turns that score into an abort decision without changing the executor. Trained on GPT-family SWE-rebench trajectories and evaluated on SWE-bench Verified and a SWE-bench Pro subset with GPT and Claude executors, StopLoss is compared against six general-purpose judges reading identical windows. Outcome supervision lifts failure prediction from near chance to a useful signal at the first check, at one to two orders of magnitude lower cost per check than any tested judge. This translates into savings: relative to completing every task, StopLoss lowers cost per retained solve by 4.5\% on Verified while preserving 99\% of successful runs, and by up to 31\% on the costlier Pro subset, reaching 48\% with a Qwen3-8B backbone, and in all four settings its cost per retained solve is below that of every judge policy. Outcome-trained small monitors can make failure recognition cheap enough for early stopping to pay for itself.

0.35

Toward Agentic Retrieval-Augmented Generation for Enterprise Retail Workflows over Heterogeneous Data Sources

cs.AI cs.IR ·2026-10-18 ·Jeonghwan Choi, Jibin Hwang, Hayoon Park, Hyeonjae Cheon, Hwanjun Song
benchmarkcode/data
Abstract · Preprint. Under review

Enterprise retail workflows require agents to chain evidence across heterogeneous sources, from structured product and order records to unstructured inquiries and reviews, yet existing agentic RAG benchmarks rarely capture this challenge. To enable controlled evaluation of such workflows, we introduce RetailBridge, a compositional retail benchmark built from persona-driven simulations. Its questions are constructed by linking entities and intermediate answers across heterogeneous sources, requiring agents to identify the relevant sources, carry evidence between them, and produce a grounded answer. To systematically evaluate LLMs on these retail workflows, we develop RetailRAG, a retail-aware framework equipped with auto-schema briefing, source-routed plan chaining, and diagnostic plan revision. RetailRAG consistently outperforms existing agentic RAG baselines across all evaluation axes, serving as a strong reference system for assessing LLM agents on retail workflows. On RetailBridge, evaluating ten LLM backbones with RetailRAG reveals multi-source chaining as the dominant bottleneck across providers and model scales, suggesting that reliable retail agents require explicit cross-source orchestration rather than scaling alone. Code and data are publicly available at \url{https://github.com/DISL-Lab/RetailBridge}.

0.35

Emergent Sparsity: How Many Effective Variables Survive Coarse-Graining in AlphaEarth Embeddings?

cs.LG cond-mat.stat-mechphysics.geo-ph ·2026-10-18 ·Johannes Hirn
publishedunusual
Abstract · Camera-ready version of a paper accepted at the 2nd Workshop on Advances in Representation Learning for Earth Observation (REO), NeurIPS 202…

AlphaEarth encodes each 10 m land pixel as a 64-component vector intended to support diverse global mapping tasks. Yet many scientific applications work at coarser spatial supports, for which fewer than 64 effective variables may suffice. We ask whether these vectors can be reorganized into ranked global nonlinear coordinates, how many suffice at a chosen reconstruction accuracy, and whether that number decreases with spatial scale. We thus describe effective dimension as a two-control surface over tolerance and spatial scale, rather than one intrinsic number. At each scale, a $β$-VAE scan with a common architecture produces a rank--distortion envelope, giving the smallest retained count achieved within the scanned family at each reconstruction tolerance. At 3% validation tolerance, this number falls from 24 to 7 when the scale is increased from 10 m to 78.9 km, whereas the corresponding PCA rank falls only from 46 to 38. Repeating the analysis on unit-norm targets preserves this trend.

0.35

World Models, Trees and Foundation Models for Retail Forecasting: Evidence from Four Public Data Sets

cs.LG cs.AI ·2026-10-18 ·Karthik Iyer, Madan Dabbeeru
code/data
Abstract · 37 pages, 18 figures, 26 tables. Code: https://github.com/karthik-iyer-au/retail-world-models, archived at doi:10.5281/zenodo.23448056

A retailer makes different decisions at different grains of the product, store and time hierarchy: store orders for an item in a store, supplier funding for an item across stores, category performance by store, and the financial plan for the chain. Which forecasting model to use depends on the grain. This paper compares models at each grain on four public data sets (M5, six Dominick's categories, Breakfast at the Frat, and Corporación Favorita, the last kept out of all development), over rolling forecast origins with paired tests and model confidence sets. It covers price and promotion through the plan inputs these data record, not product, assortment or place. The models are a world model that carries each item's state forward under the price and promotion plan, the same network without that state, LightGBM, the Temporal Fusion Transformer, TimesFM, Chronos-Bolt, Timestral, seasonal naive, and an equal-weight average of the world model, LightGBM and TimesFM given the plan. The answer changed with the grain. For an item by store and an item across stores, the average had the lowest error on seven and eight of nine data sets, and the 90% confidence set held a median of 1 and 2 models. For a category by store, a category and the chain total, the lowest-error model differed between data sets and a median of 4 of nine models were not separated by the four to six windows available. For category-by-store planning cost, the average or LightGBM was lowest on seven of eight data sets, and the world model was lowest only on M5; it was behind LightGBM item by store on six of eight. Every trained model used the plan, and on the weekly promotion-driven data pretrained models were less accurate without it. These results hold for these data sets; a forecast solved across all four Ps, or a retailer's own data, may give different results. Every model, test and change that lost is reported.

0.34

Consistent LLM Call Handling Across Java Frameworks: A Reference Implementation and Benchmark

cs.LG ·2026-10-18 ·Oleksandr Klymenko
benchmarkcode/data
Abstract · 7 pages, 2 figures, 3 tables, 1 listing. Code and data: https://github.com/alxkm/llm-call-guard

Java applications reach large language models (LLMs) through an integration framework, and rely on it to handle a failing provider: to retry what can be retried and nothing else, to wait as long as the provider asks, never to present a truncated answer as complete, and to record each failure in the application's metrics. We state these expectations as a six-clause reliability contract and measure Spring AI 2.0.1 and LangChain4j 1.21.0 against it with a scripted OpenAI-compatible provider: fifteen failure scenarios, synchronous and streaming, three repetitions. LangChain4j meets it in 47 of 77 applicable cases with its defaults and in 51 of 83 when configured explicitly, Spring AI in 72 of 77 and 76 of 83; both deliver a stream that ended without its terminator as a complete answer. We then place one implementation of the contract beneath both frameworks, at the HTTP boundary - an OkHttp interceptor for Spring AI, an HttpClient for LangChain4j, and a java.net.http transport for code without a framework - sharing a 725-line core that depends only on the JDK and the Micrometer Observation API. With it, all three clients behave identically and meet the contract in 90 of 91 cases; the remaining case is a policy trade-off between retrying a hung provider and waiting for a slow first token. Against a local model server the layer rejected none of 700 calls, and it adds a median of 32-217 microseconds to a call.

0.34

Neural Networks with Differentiable Belief-Function Layers

cs.AI ·2026-10-18 ·Antonin Sulc
published
Abstract · 10 pages, 2 figures, BELIEF 2026

Yager's 1987 rule of combination $\boxplus$ was proposed on epistemic grounds: under high conflict, mass should be redistributed to ignorance rather than renormalised onto the class simplex. We observe that backpropagation needs the same rule: Dempster's normaliser $(1{-}κ)^{-1}$, where $κ$ is the conflict mass, puts a $(1{-}κ)^{-2}$ term in the gradient that diverges exactly when evidence sources disagree, while Yager's $\boxplus$ is bilinear and has no such term. Building on this, we construct a deep evidential pipeline in which every mass in the chain from heads to output is a valid Dempster--Shafer mass function, combined by Yager's rule. The output mass on the ignorance set, $m(Θ)$, serves as the network's ``I do not know'' signal. The network is fit to hard labels and never sees an annotator distribution, yet $m(Θ)$ orders human disagreement in two modalities, visual and linguistic. Dempster's rises too and ranks contested inputs about as well, but its normaliser divides the conflict out: Yager's vacuity is $9\times$ larger on the most contested CIFAR-10H images, so only Yager's is readable as an amount of ignorance without recalibration. For out-of-distribution detection it is competitive with strong single-pass baselines, while the clamped Dempster variant that scores higher becomes gradient-unstable once the clamp is removed.

0.34

A Distribution-Free Separation Between Learned and Fixed Features

cs.LG stat.ML ·2026-10-18 ·Daniel Barzilai
hot direction
Abstract

The success of neural networks is largely attributed to their ability to learn useful representations from training data. However, known results that prove a computational advantage of learned features over fixed ones are restricted to specific input distributions. We settle the COLT open question of Feldman, Kamath, and Srebro (2026) by showing that this advantage can persist in the distribution-free setting. We construct a family of hypothesis classes that can be efficiently learned by a natural gradient-based algorithm on a two-layer ReLU network, but cannot be represented by linear predictors over any fixed feature map with polynomial dimension (and runtime). Specifically, for any sufficiently small error and any sufficiently large runtime budget R (depending on the error), there exists a hypothesis class that the neural network learns under any input distribution to the desired expected classification error in time at most R, while the dimension complexity (governing the dimension needed by linear predictors) is superpolynomial in R.

0.33

Policy Is Not Capability: Some Safety Benchmarks Measure Which Definition of "Unsafe" You Trained On

cs.LG cs.AIcs.CLcs.CR ·2026-10-18 ·Vahid Jebraeeli, Jonathon Romero, Yijin Wang, Adrian Chan, Jianbang Zhang
publishedbenchmarkposition
Abstract · Accepted as a poster at the NeurIPS 2026 Workshop TAE (Trust-AI-Eval): Can We Trust AI Evaluation?, Sydney. 20 pages, 5 figures

A guardrail classifier is reported as if unsafe were a property of text. It is not: it is a property of a policy, and the policy enters the model through its labels. We exploit an accident of an industrial deployment in which the same ~53k prompts were labelled twice, under two operational definitions ("is this a request for harmful content?" and "would a moderator block this?"), and used to train ten guardrail arms sharing architecture, codebase and optimiser family, at two model scales. First, the guardrail's own training corpus already contained the benchmarks: JailbreakBench-benign is 88% and OR-Bench-hard 82% present in the training pools, two over-refusal benchmarks, the kind a guardrail reports to argue it does not over-block. Everything below is decontaminated. Second, safety benchmarks stratify sharply by how much of their score is decided by the label definition rather than by the model: with architecture and optimiser fixed, the range across the ten arms is 0.99 on HarmBench-copyright and 0.043 on WildGuard-benign, a 23x spread. Third, a benchmark score can be moved across its entire range by 0.03% of the parameters. Across these arms HarmBench-copyright spans 0-100 and OR-Bench-hard over-refusal 3.4-56.4%, and the four arms whose last supervision signal carries a given policy occupy one extreme of both (exact rank p=0.005, or 0.04-0.14 at worst), at unchanged detection capability. For a distilled student the teacher's soft labels do this even when every hard label carries the other policy, so teacher provenance is an unrecorded confound in any released checkpoint. Fourth, validation accuracy on the one split that grades all twenty arms tracks agreement with the annotator's policy (rho=+0.74), not detection capability (rho=+0.34). We argue that guardrail evaluations should report contamination, a seed floor and the definition they were graded against as first-class variables.

0.33

Structured Relevance Determination and Channel Additive Gaussian Processes for Image Data

cs.LG ·2026-10-18 ·Harald Roßels, Ludovico Scarton
code/data
Abstract · Code: https://github.com/hrosse/structrd-caddgp

We introduce radial basis function (RBF) kernels for Gaussian processes that use structured hyperparameter sharing for image-valued inputs. For high-dimensional inputs such as images, a fully anisotropic RBF kernel assigns a separate lengthscale to each input dimension, which can introduce many hyperparameters and may lead to overfitting and poorly calibrated predictive uncertainty when data are limited. We propose structured relevance determination by tying lengthscales across groups of adjacent pixels and within each colour channel. This yields patch-wise and channel-wise constrained anisotropic RBF parametrisations that reduce the number of independently learned lengthscales. We further introduce channel-additive Gaussian processes for multi-channel images and combine them with structured relevance determination, resulting in a parameter-efficient decomposition of the latent function into components defined on channel subsets, enabling some non-local generalisation. The proposed kernels are evaluated on MNIST and a compact binary aircraft-detection dataset of low-resolution RGB satellite image chips. We also show how channel-wise structured relevance determination can be incorporated into convolutional Gaussian processes. We observe improved predictive performance compared to isotropic and fully anisotropic RBF baselines.

0.33

ATEST: Adaptive Resource-Aware Test-Time Scaling for LLM Inference

cs.LG cs.AIcs.CLcs.DC+1 ·2026-10-18 ·Mobina Kashaniyan, Ali Jannesari
Abstract · 12 pages, 4 figures, 8 tables

Test-time scaling (TTS) improves large language model (LLM) reasoning by sampling multiple answers, but fixed sampling budgets ignore both query difficulty and the changing energy cost of inference under GPU contention. We measure up to a 4.1-fold increase in energy for the same four-sample query as GPU load rises. We present ATEST, a resource-aware TTS framework that jointly considers the expected value of additional samples and their current GPU cost. ATEST-Online, its default mode, requires no offline calibration: it incrementally generates samples, estimates the stability of the plurality answer, monitors live GPU conditions, and learns step energy cost online to decide when to stop. ATEST-Calibrated instead selects a per-query budget using offline accuracy and cost models, favoring deployments with stable loads. We evaluate both modes on four reasoning and question-answering datasets using Phi-3-mini on an NVIDIA A100 under controlled varying load, and test transfer to an NVIDIA L40S and Qwen2.5-3B. Across the four A100 datasets, ATEST-Online answers every query without exceeding the energy budget, while achieving the highest or statistically tied accuracy and reducing sampling as load increases. On GSM8K at high load, it achieves 60.0% accuracy with no deferrals, versus 44.0% accuracy and 42.4% deferrals for ATEST-Calibrated. Ablations show that online cost learning drives adaptation to load. Including offline profiling, ATEST-Online uses 42 times less GPU time and 48 times less energy than ATEST-Calibrated for the evaluated workload. Online control trades higher tail latency and lower throughput on short tasks for robustness under varying load; offline calibration remains competitive under stable conditions.

0.33

Polynomial Learning of Positive ReLU Networks under Gaussian Inputs

cs.LG ·2026-10-18 ·Zhijie Zheng
Abstract · 32 pages; expanded to include agnostic learning guarantees

We study positive, bias-free, one-hidden-layer ReLU networks under standard Gaussian inputs. We obtain algorithms with polynomial sample and real-arithmetic complexity, without separation or conditioning assumptions on the hidden directions. When the conditional label mean is a width-$k$ network with total output weight at most one, the learner achieves squared prediction error $\varepsilon$ with output width at most $k+\min\{k,d\}$ and complexity linear in one plus the residual second moment. For arbitrary labels with bounded second moment, we obtain a positive ReLU network of width $O(k\log(2/\varepsilon))$. Its risk is bounded by the irreducible noise variance plus $Ck^3\log^{9/2}(2k/\varepsilon)$ times the squared distance of the conditional mean to the target class, plus $\varepsilon$. The proofs combine implicit Hermite moments, orthogonal tensor-train compression, and a direction decoder whose probability of returning a direction far from every atom is bounded linearly in the compression and estimation errors. Exact Gaussian localization refines the resulting direction covers. In the agnostic setting, a cone-constrained convex program with a scale-dependent penalty produces a positive network; in the conditional-mean setting, exact sparsification yields the smaller width. We also establish projective direction recovery and confidence amplification with sample complexity proportional to the logarithm of the reciprocal failure probability.

0.32

Mining the Self-Rewriting System: A Vision for Data Mining on AI-Native Infrastructure

cs.LG ·2026-10-18 ·Mayank Darbari, Chiman Salavati, Liang Wu, Dawei Wang, Jiong Zhang +3 more
top venue
Abstract · Accepted to the Blue Sky Track of the IEEE International Conference on Data Mining (ICDM)

Classical data mining treats data as generated outside the learner. Emerging AI-native infrastructure invalidates this assumption: as agents observe system state, act, update retrieval stores and models, and promote new artifacts into the serving path, their own decisions reshape the future logs, traces, corpora, and evaluation records used for later training and release. We call this new research object a self-rewriting system and the resulting system-induced distribution shift endogenous drift. Our central, falsifiable claim is that cross-surface causes are not identifiable in general from marginal per-surface monitoring: different coupled failures can produce the same individual signals while differing in their joint behavior, requiring joint analysis for reliable attribution. This agenda comprises four foundational research problems: fusing heterogeneous drift signals into calibrated gating and attribution (C1); mining agent traces whose activity alphabet evolves (C2); maintaining knowledge fabrics through online reindexing, faithfulness, and compression decisions (C3); and safely promoting concurrent learning streams through one shared evaluation gate (C4). These remain fragmented across concept drift, performative prediction, process mining, streaming retrieval-augmented generation, and MLOps, without a unified data-mining agenda. We outline public benchmarks and evaluation criteria, using AI-native telecom only as a motivating exemplar with analogues in finance, software CI/CD, and clinical decision support.

0.32

One Skill Takes Over: Clashing Agent Skills and a One-File Fix

cs.AI ·2026-10-18 ·Parsa Razmara, Seyedarmin Azizi, Ardalan Aryashad
code/data
Abstract

Agent Skills are instruction files that tell a coding agent how to do one job. Users install skills from many authors, so two skills for the same job can be active at the same time. We ask what a model does when both skill files are in its context: does it combine their compatible rules, or follow one skill and drop the other? We build SkillClash, 16 pairs of same-job skills in which code checks every rule and a reference output shows the best result that can be reached. On seven models from five families, from 4B open models to Qwen3-235B and GPT-4.1, a skill's rules are followed 93 to 99% of the time when it is loaded alone and 61 to 62% when a same-job skill is loaded next to it. Most outputs show takeover: every rule that only one skill asks for is followed, and every rule that only the other skill asks for is dropped. The worked examples inside the skills drive this. When each skill file instead ends with one example that follows both skills, takeover falls to 1 to 15% on the six models we tested. Telling the model which skill has priority does not bring the dropped skill back, and telling it to apply both skills brings the dropped skill back only on the two strong models and leaves the interacting rules broken. One merged skill file written by the strongest model we used, Claude Opus 5.5, lets all seven models follow 96 to 99% of the achievable rules. Most of its gain over a plain list of both skills' rules comes from its one worked example, and such an example helps only when it is correct. Code and data: https://github.com/prazmara/skillclash

0.32

Spectral Adaptive Graph Transformer for Efficient and Scalable Traffic Forecasting

cs.AI cs.LG ·2026-10-18 ·Amirhossein Ghaffari, Susanna Pirttikangas, Ekaterina Gilman
benchmarkcode/data
Abstract · 16 pages, 2 figures, 9 tables. Code: https://github.com/Ahghaffari/sagt

Traffic forecasting models based on spectral graph convolution offer an interpretable account of how information is mixed between sensors, but they rely on an eigendecomposition of the graph Laplacian. It costs cubic time, is fixed by the road network, and provides a symmetric operator that cannot represent the directed propagation of traffic congestion. Attention-based spatio-temporal transformers remove these limitations but generate a dense sensor-by-sensor attention map at each layer and time step, so memory, rather than accuracy, becomes the binding constraint on large networks. We introduce the spectral adaptive graph transformer, whose spatial operator is learned rather than derived from the Laplacian, directed rather than symmetric, and conditioned on the current traffic state rather than a static state, at a cost linear in the number of sensors. A sparse road graph diffusion branch provides the strictly local propagation that a low-rank operator cannot express, and a zero-initialized gate learns the balance between the two. Before spatial mixing, a unit-mean-gain frequency-bin attention module reweights each sensor's short-term periodicities according to its identity and current context. Across six public benchmarks spanning 170 to 1843 sensors, evaluated over three random seeds, the proposed model attains the lowest error of any single end-to-end model on one benchmark and remains competitive with the strongest current attention models on the others, while using about a quarter of their memory and training about twice as fast.

0.32

Estimating Causal Bounds from Observational Data With Neural Generative Models

cs.LG stat.ME ·2026-10-18 ·Max Zhu, Srivathsan Amruth, Sina Akbari, Qingyuan Zhao
benchmarkcode/data
Abstract · 23 pages, 3 figures, 3 tables

Identifying causal bounds from discrete observational data is a fundamental problem in causal inference. While exact analytical bounds exist for certain graphs, this is not generally the case. In this paper, we introduce a framework for estimating such bounds using neural generative models together with a new extrapolation procedure to estimate these bounds more accurately. Across multiple causal inference settings, it produces bounds that are consistently closer to the true bounds than those obtained with existing general neural approaches. We validate the framework on a range of benchmark problems for which exact bounds are available.

0.31

From Clinical Risk Prediction to Operational Decisions: An Uncertainty-Aware Framework for ICU Staffing

cs.AI cs.LG ·2026-10-18 ·Jiayan Yang, Ad Feelders, Arno Siebes
published
Abstract · Accepted at ECML PKDD 2026, Applied Data Science Track

Advances in electronic health records (EHR) and large-scale critical care databases have enabled the development of machine learning models for the early detection of patient deterioration. While these models often achieve strong predictive performance for outcomes such as sepsis and mortality, most studies evaluate them purely on discrimination metrics. This leaves a critical gap: how to systematically translate probabilistic clinical risk estimates into downstream operational decisions under real-world constraints, such as shift rigidity and hiring delays. This study proposes an uncertainty-aware anticipatory staffing framework that bridges short-term patient risk prediction and staff planning without altering the upstream predictive model. Instead of relying on naive deterministic aggregation or fixed safety buffers, the framework explicitly models dynamic predictive uncertainty and evaluates extreme shortage exposure using Conditional Value-at-Risk (CVaR). These aggregated, risk-calibrated demand signals are then embedded into an hourly rolling-horizon scheduling policy to generate delay-constrained staffing decisions. Evaluated on realistic trajectories from the MIMIC-IV database, the proposed framework demonstrates a favorable cost-shortage trade-off compared to deterministic and purely reactive baselines. Pareto frontier analysis illustrates how uncertainty-aware demand construction changes resource allocation under varying safety preferences, while stress testing confirms the framework's operational robustness even when subjected to systematic predictive bias in the form of calibration drift. By formalizing this translation layer, this work provides a principled, transparent pathway for safely integrating probabilistic AI outputs into high-stakes healthcare operational systems.

0.31

Sampling and Curvature in First-Exit Comparisons for Momentum SGD

cs.LG ·2026-10-18 ·Haiyi Li
Abstract · 84 pages, including references and appendices

First-exit prediction for momentum SGD depends on sampling information that pointwise gradient laws omit. Strongly convex finite pools with identical pointwise laws and matched iid momentum dynamics can have different reshuffled exits; even analytic pools with uniformly close pointwise laws can have separated exit probabilities. For persistent sampling in one dimension, we establish an explicit $O(h)$ mean-exit comparison with a finite-rate switching process under nonvanishing component velocities and step-size conditions; iid sampling instead averages to an ODE. Retaining realized minibatches and sampled curvature yields finite-window exit brackets beyond this regime. In 32 trained ridge-logistic runs, 29 meet prespecified targets with rounding-certified brackets; an enclosure control explains most of the width advantage. Exact-real certificates establish finite means at the original tested radii for three ResNet-18 checkpoints and give local finite-mean conditions for tanh and ViT networks; their upper bounds remain loose, and original-radius ViT mean guarantees remain unresolved despite 128 fresh continuation paths. A two-network replication gives 32/32 successful seeds in each of four finite-path persistence comparisons. Independent equal-length controls retain slower exits after refresh resets than without resets, but the refresh-versus-random ordering reverses with update length at both networks. These interventions do not causally decompose raw momentum SGD. Neural step-size scaling does not validate the switching limit.

0.31

The Hidden Look-Ahead: Point-in-Time Leakage in Electricity Market Data

cs.LG cs.CEstat.AP ·2026-10-18 ·Abdelkrim Alfalah
positioncode/data
Abstract · 13 pages, 4 figures, 7 tables

Empirical studies of electricity prices routinely state that their models use only information available at the time of the trading decision. For weather-driven forecasts drawn from commercial archives, this condition is harder to meet than it appears, and whether it holds can change over time. We identify three timing pitfalls. A forecast run is labelled by the issue time of its weather model, but becomes available hours later and keeps changing after publication. Auction gates follow local clock time, while archives use UTC. Outturn series are revised after publication without earlier values being kept. Using the timestamps of 128,656 forecast runs from ten weather-model configurations, matched to the day-ahead and first intraday auction gates in Great Britain over 2022-2026, we find that selecting "the latest run issued before the gate" picks a run not yet published at the gate on more than 99% of delivery days for six of the ten configurations at one of the two gates. For GFS at the day-ahead gate, the share is 0.6% in winter and 99.8% in summer. The exposure also shifts as publication delays drift, and an action margin of a few minutes can cost a full six-hour forecast cycle. The wrongly selected inputs differ materially from those available at the gate, but their apparent accuracy gain ranges from -9% to +24% across variables, models, gates and seasons, so its effect cannot be assumed and must be measured. We argue that point-in-time correctness is a dynamic property of the data pipeline, and propose a protocol that audits it from timestamps before any model is fitted.

0.30

Strictly Replicable Realizable Learning: Unknown Marginals, Finite-Class Interpolation, and the Cost of Shared Randomness

cs.LG stat.ML ·2026-10-18 ·Guangjian Zhang
theoryhot direction
Abstract · 68 pages

Strict replicability requires two independent training samples, processed with the same random seed, to produce exactly the same hypothesis. We study realizable Boolean learning under unknown and known instance marginals. For unknown marginals, we obtain the worst-case sample complexity $Θ(((\log M)^2/ρ^2+\log(1/δ))/\varepsilon)$ for classes of cardinality at most $M$, throughout a fixed small-parameter range. This resolves the distribution-free finite-class sample-complexity question up to universal constants and rules out a general $ρ^{-1}$ dependence. The upper bound uses a truncated-power distribution with a hard empirical-error budget. For known marginals, a control-variate learner uses $O(((1+\log M)^{3/2}/ρ+\log(1/δ))/\varepsilon)$ samples, and explicit uniform-marginal instances establish the matching class-size exponent. On a product of $d$ finite threshold chains of length $k$, we determine the exact rate $Θ((dk\log(1+\sqrt d/(kρ))+\log(1/δ))/\varepsilon)$. We also characterize deterministic learning, the very-small-$ρ$ regime for arbitrary finite classes, and every cardinality-replicability regime for finite classes of VC dimension one. A VC-covering-number bound yields the fixed-accuracy exponent $3/2$ independently of cardinality. Finally, we separate statistical complexity from implementation costs: we prove a sharp fixed-block trade-off, short-seed existence bounds, and a shared rejection-filter reduction from optimal learning of the full Boolean class to efficient public sampling of ordinary Bernoulli product distributions. Detailed proofs include the geometric lower bounds, the discrete posterior-kernel comparison, and the finite-bit constructions.

0.30

The Optimal Sample Complexity of Pointwise Replicable Agnostic Learning

cs.LG ·2026-10-18 ·Guangjian Zhang
theory
Abstract · 92 pages

Pointwise replicability asks that two independently trained predictors, using the same random seed, agree at each fixed query with high probability. We determine its optimal agnostic sample complexity on finite and countable domains. For every nonempty binary class of finite VC dimension $d\ge1$, the required number of independent examples is $Θ\!\left(\frac{d}{ρ^2α^2}+\frac{\log(1/β)}{α^2}\right)$, uniformly over independent accuracy, failure, and disagreement parameters $α,β,ρ\in(0,1/10)$. The learner returns one binary function, may be improper, and uses unrestricted shared randomness. Its sample budget is fixed before observing data, and the replicability guarantee includes queries of zero marginal mass. The main construction combines an ensemble of empirical-risk minimizers with a smooth, signal-dependent risk decomposition. A single shared phase selects the final binary predictor. Record times certify its weak-signal risk, while a uniform bound protects the strong-signal risk at that same selected time. A phase-comparison theorem controls the actual shared-seed execution under simultaneous changes in the marginals and risk coefficients. A rare repair step then makes the confidence cost additive. Together with matching lower bounds, this removes the remaining logarithmic overhead and parameter restrictions in the general countable-domain sample-complexity problem. This extended version also retains the earlier explicit constructions and their proofs, including proper learning for VC-one classes, the full finite-atom class, a faster realizable bound, and counterexamples that separate restrictions of particular algorithms from restrictions of the learning problem.

0.30

Auto-MD-bench: Evaluating Autoresearch Agents for Real-World Medical Tasks

cs.LG cs.AI ·2026-10-18 ·Shenqi Zhou, Qiran Zou, Wenhao Zhao, Zhenyu Liao, Dianbo Liu
benchmarkcode/data
Abstract · 24 pages, 5 figures, 22 tables. Code: https://github.com/Edwin0080/Auto-MD-bench ; data: https://huggingface.co/datasets/Edwin0080/Auto-MD-b…

Autonomous agents are increasingly capable of developing machine learning (ML) models end to end. In medical practice, however, strong performance on the target metric does not fully characterize how good the models are in real-world use cases. For example, medical ML models should also provide well-calibrated risk estimates and remain robust across patient populations. Yet autonomous model development is typically guided by task objectives that emphasize simple predictive discrimination (e.g. AUROC). This raises a key question: are these objectives sufficient to guide agents toward models that meet the broader requirements of real-world medical use? In this study, we introduce Auto-MD-bench, a benchmark of 15 real-world medical tasks from 10 clinical datasets, spanning survival analysis, classification and sequential prediction, with external validation across cohorts, hospitals and time periods. We evaluate four agents: AIDE, AI Scientist v2, Codex, and Autoresearch. The strongest agents achieve competitive primary scores, yet their predicted risks are less calibrated than those of the simple baseline models, which could lead to severe medical accidents. Moreover, models with nearly identical primary scores can exhibit markedly different calibration, while performance varies substantially across population shifts. These findings show that primary task performance alone is insufficient to assess autonomous clinical modeling and multiple different metrics widely used in healthcare practice need to be taken into consideration.

0.29

CoworkEnv: Synthesizing Training Environments for Long-Horizon Cowork Agents

cs.AI cs.CL ·2026-10-18 ·Zhongwei Xie, Yiming Du, Jiaxin Bai, Baixuan Xu, Chunyang Li +8 more
community pick
Abstract · work in progress

AI coworkers must coordinate tools, communicate across turns, and act on changing workspace state while respecting workflow constraints. Constructing training experience for these capabilities requires coherent requests, service records, tool interfaces, and grading checks. We present CoworkEnv, a capability-guided framework for synthesizing executable training environments for AI coworkers. Workflow requirements guide task construction, and programmatic validation with feedback repair checks generated specifications and grading behavior. Teacher executions yield curated demonstrations for supervised fine-tuning (SFT), while student interactions provide rollouts and rewards for reinforcement learning (RL). Using these resources, we train compact PocketWorker models. For Qwen3-VL-8B-Instruct, SFT raises the mean score from 46.6 to 64.1 on Claw-Eval, including a 31.9-point gain on Multi-turn tasks. An SFT-only extension to ClawMark demonstrates applicability across coworker benchmarks. We further compare RL training dynamics across configurations that vary reward processing and credit assignment. Improving training rewards provide preliminary evidence that the constructed environments support policy optimization.

0.29

Learning When to Abort in Transformer Planning: A Mechanistic Study

cs.LG cs.AIcs.CL ·2026-10-18 ·Shi Feng, Qimin Zhong, Siwei Wang, Yifei Shen, Shang-Hua Teng +2 more
theory
Abstract

Planning does not by itself imply the ability to recognise errors, but recognising errors can improve planning capacity. In this paper, we study this phenomenon in Transformer planning. To analyze the inner mechanism, we follow the approach of Wang et al. (2024) and abstract planning as path-finding in graphs. In this setting, we train the model to detect the error of invalid transitions, i.e., after the model generates a next node that is not adjacent with the current node, the model should then generate a special abort token x. For expressiveness, we construct a Transformer with feed-forward width $O(|V|)$ for any fixed graph with node set $V$. It can detect all possible invalid transitions, and select a valid next node if the prefix is valid. We then propose a worst-case lower bound on feed-forward width, which makes the detector's linear width near-optimal up to logarithmic factors. In trained models, we demonstrate that invalid-transition detection increases previous-node attention, and most of the valid--invalid separation locates in the feed-forward network's output. This accords with our theoretical insights. Finally, we use this error detection to improve the model's performance: at inference, generating the abort token x will directly cause an external procedure to discard the attempt and restart from the original prompt. Experiments show that error detection combined with retry improves model's planning reliability on graphs, Countdown, and mazes.

0.29

LowNet: Looped Wavelet Neural Networks with 24 Learnable Parameters for Solving Poisson Equations

cs.AI cs.CVcs.LGeess.IV+2 ·2026-10-18 ·Yuanhao Gong
Abstract

Poisson equations frequently appear in various fields, such as mathematics, physics and informatics. Solving Poisson equation, however, is challenging. Traditional methods are not data-specific, although have high certainty. They require a large number of iterations due to the CFL numerical stability. In contrast, deep learning methods are data-driven and more efficient, but require accurate training data and suffer from heavy computation. To tackle these issues, we propose a wavelet-inspired Poisson solver with 24 trainable scalars. The proposed network has several advantages. First, it is compact. It has four independent $5\times5$ kernels, each represented by six parameters, leading to 24 scalars in total. Such compact format can be deployed on most devices. Second, thanks to the loop design, it achieves higher accuracy than previous methods. A residual loop reuses the wavelet block without increasing its parameters. Numerical experiments confirm that such loop indeed improves the solving accuracy. On 10000 images, one, two, and three loops of the wavelet block lead to PSNRs of $47.27$, $74.14$, and $100.74$ dB, respectively. Third, the proposed method is computationally fast. Thanks to the wavelet transform, the proposed method has linear computation complexity, leading to high computation performance. In short, we propose a novel looped wavelet neural network with compactness, accuracy and efficiency for solving Poisson equations. Our method can be applied in various fields, such as computational fluid dynamics, image reconstruction, biomedical engineering, electromagnetic field simulation.

0.27

Indistinguishable Agents: Training Agent Privacy as Twin-World Indistinguishability

cs.LG cs.CL ·2026-10-18 ·Jiaming He, Guanyu Hou, Yi Yu, Xudong Jiang
theory
Abstract

LLM agents search, shop and send messages for their users, and every action is seen by some party. Existing privacy training rewards agents for not mentioning protected items, yet an observer can infer a private fact from what the agent does. Since an observer learns a fact only if its view changes with the fact, we propose \methodfull (\method), which trains on twin worlds that differ only in the private fact. A frozen distinguisher scores each observer's view by its log-likelihood ratio for the true world, and paired GRPO trains the agent on task success minus this ratio. We prove that this objective controls the Bayes success of every observer on every channel. On shopping and messaging tasks \method removes the leakage that disclosure judges and query-level rewards miss, while the agent still completes its tasks.

0.27

MAMath-Bench: Process-Aware and Robust Evaluation of Multimodal Multi-Agent Mathematical Reasoning

cs.AI ·2026-10-18 ·Xinping Wang, Jie Ma, Rui Xing, Wensong Guo, Xiayi Zhang +4 more
benchmarkcode/data
Abstract

Mathematical reasoning is a fundamental capability for educational applications of Large Language Models (LLMs), where reliable problem solving requires not only correct answers but also coherent and verifiable reasoning processes. Multi-agent systems (MASs) have shown growing potential in this domain by leveraging collaboration and verification among multiple LLM agents. However, existing evaluations primarily focus on the correctness of final answers in text-only settings, offering limited insight into multimodal reasoning capabilities, the reliability of intermediate reasoning steps, and robustness against reasoning errors. To address these limitations, we introduce MAMath-Bench, a process-aware benchmark designed to evaluate the multimodal mathematical reasoning capabilities of MASs. MAMath-Bench comprises 1,003 problems spanning 11 mathematical categories, multiple difficulty levels, and several modalities, including text, images, and audio. It features fine-grained milestone annotations to evaluate systems across three dimensions: problem-solving correctness, reasoning-process reliability, and robustness to adversarial reasoning-chain perturbations. Our evaluation yields three key findings. First, multi-agent collaboration does not consistently improve final-answer accuracy over strong single LLMs. Second, collaboration can substantially enhance the coverage of intermediate reasoning, but such gains may come at the cost of increased reasoning redundancy. Third, current systems remain vulnerable to adversarial reasoning errors, particularly those involving subtle structural perturbations. The dataset is available at https://github.com/mira-ai-lab/AgentMath.

0.27

Dynamic System Inspired Continuous Large Language Model for AI4Science Tasks

cs.LG cs.CL ·2026-10-18 ·Zhaoxin Fan, Zheng Zhu, Yuqing Ma, Faguo Wu, Jian Yang +1 more
Abstract

Scientific foundation models increasingly operate on continuous phenomena, yet still use discrete residual computation. We ask whether pretrained residual transformations can support additional latent computation without expanding the backbone. To test this possibility, we introduce Continuum-LM, which reuses these transformations as dynamical operators in a continuous-depth language model. Its hidden states evolve through an ordinary differential equation evaluated by higher-order numerical integration, with solver order and step size controlling computation. Across eight scientific tasks, Continuum-LM trains stably and improves mean exact-match accuracy by 7.77 and 10.96 percentage points over discrete counterparts at the 0.6B and 1.7B scales, respectively. At 0.6B, it also outperforms an Ouro-style recurrent control under matched parameter and residual-function evaluation budgets. Together, these results support extensible computational depth through parameter reuse, identifying continuous hidden-state evolution as a practical design dimension for scientific language models beyond parameter scaling.

0.27

Steered Not Stirred: Effective Reward Tilting using Surrogate Reward Mixing

cs.LG ·2026-10-18 ·Alec Sargood, Olivia Gallup, Georg Maierhofer, Noam Ghenassia, Paul Duckworth +2 more
Abstract

Continuous-time generative models, from diffusion models, flow matching, and now flow maps, owe their practical value to high-fidelity generation and steerability. In many applications, a ground-truth latent ``oracle'' reward $R^*$ is unavailable or expensive to evaluate, e.g., human judgments or experimental measurements. Although the oracle reward is not directly observable, it can often be approximated through multiple cheaper surrogates that, individually, provide incomplete characterizations but collectively reveal substantial information about the downstream task. In this paper, we present a method for composing such surrogates to steer a pretrained model toward the distribution tilted by an unavailable $R^*$. Given a small feedback set of oracle rankings or numerical rewards, we introduce Mixing And Reward Tilting for INference-time Improvement (MARTINI), which learns a surrogate composition for plug-and-play use with standard inference-time steering methods. Our theory characterizes surrogate information limits and relates population recovery error to ideal-target KL at small tilt strength. For preferences, the bound additionally accounts for an irreducible mismatch between the Plackett-Luce population target and the conditional mean. Experiments demonstrate improved oracle-tilt fidelity in language generation and ImageNet alignment, and improved joint pose accuracy and validity in protein-ligand cofolding, without inference-time oracle queries.

0.27

Length-Independent Stability and $L_2$-Gain Certificates for Chains of Delayed Neural Modules

cs.NE ·2026-10-18 ·Ömer Faruk İşcanlı
unusual
Abstract · 11 pages, 4 figures, 2 tables. Submitted to IEEE Transactions on Systems, Man, and Cybernetics: Systems

Networks of repeated neural modules with delayed neighbour coupling need stability guarantees for every number of modules, but delay-dependent criteria certify one network at a time and grow with its size. This paper derives stability and $L_2$-gain certificates of length-independent size for open chains and rings of modules of two alternating types with slope-restricted activations. A local and a transmission delay act independently and are only bounded and continuous. A Lyapunov-Krasovskii functional with fixed integration limits and module-local variables that repeat with the module type reduces the certificate of every chain to one cell of two modules. Local conditions and a single linear matrix inequality of order 96 on this cell, or 104 for the gain, certify global asymptotic stability and a common $L_2$-gain bound for all open chains and all rings with an even number of at least four modules. Up to a margin, the stability inequality is also necessary for the repeated variables to certify every length. All uniform certificates, including certified coupling regions, are verified in exact rational arithmetic. In the numerical study, the certified coupling reaches 92% of an outer bound given by unstable members of the model class for short delays and 79% at the nominal delay bounds. For 16 modules at unit coupling strength, certificates of the same functional computed for that length alone raise the certified delay scale by 2% to 7% in four cases. A single stability problem with 1298 scalar variables thus replaces per-length certification of long chains at little cost.

0.27

Distributionally-Robust Policy Learning with Observational Data

cs.LG ·2026-10-18 ·Debmalya Mandal
top venue
Abstract · To appear at NeurIPS 2026. Generative AI was used only to polish writing and implement the main algorithm

Policy learning from observational data is fundamental to applications ranging from personalized medicine to welfare-program targeting, but standard approaches optimize empirical performance and may fail when the deployment distribution differs from the training distribution. We study \emph{distributionally-robust policy learning}: given observational data, we seek a policy $π\in Π$ that maximizes the worst-case policy value over an $f$-divergence ball of radius $ρ$ around the data-generating distribution. We propose an estimator built on cross-fitted doubly-robust scores for the conditional treatment effect, paired with a DRO solver that handles \texttt{KL}, $χ^2$, and \texttt{CVaR} ambiguity sets. We establish two convergence guarantees. First, the estimated policy attains the standard $\widetilde{O}(n^{-1/2})$ rate on its DRO regret, and inherits the doubly-robust property -- the rate continues to hold whenever either the outcome model or the propensity model is consistently estimated, even at slow nonparametric rates. We then propose a sample-splitting bias-correction algorithm derived from the Lagrangian dual, and show that under a Tsybakov-style margin condition on the treatment-effect, the estimator attains a fast rate of $\widetilde{O}(n^{-(1+α)/(2+α)})$ , interpolating between $\widetilde{O}(1/\sqrt{n})$ and $\widetilde{O}(1/n)$ as the margin parameter $α$ ranges over $[0,\infty)$. To our knowledge, this is the first work to incorporate a margin condition into the analysis of distributionally-robust policy learning, and the first to obtain fast rates in this setting. Empirically, on synthetic and semi-synthetic experiments calibrated to a welfare-to-work program, distributional robustness yields meaningful improvements over empirical welfare maximization when training and test distributions differ.

0.27

Probabilistic Limit Order Book Forecasting with Conditional Diffusion Models

cs.LG econ.EMq-fin.TR ·2026-10-18 ·Junyu Chen, Franklin Liu, Moshu Xu, Lijian Yang
Abstract

Modeling limit order book (LOB) dynamics is central to market forecasting and execution-strategy evaluation. While diffusion models achieve strong empirical performance, existing LOB snapshot models typically lack a learned recurrent encoder and formal statistical guarantees. We introduce LOBGrad, a conditional DDPM framework for probabilistic LOB forecasting that combines an autoregressive RNN encoder with a diffusion-based generator. In particular, we establish an excess-risk oracle inequality for the one-step DDPM noise-prediction objective under exponential mixing and regularity assumptions. We further show that, under contractive dynamics, the dependence on the conditioning-window length can be reduced to a dependence on the latent-state dimension. Experiments on BTC/USDT, GOOGL, and META show that LOBGrad improves on its encoder-free diffusion ablation across all reported queue metrics and obtains competitive results against other baselines. It also reproduces the key stylized facts of LOB microstructure, and yields reasonably calibrated queue-based predictive intervals. A stylized downstream task illustrates the potential value of the generated forecasts.

0.26

Let Them Sleep: Recursive Self-Pruning Memory for Long-Horizon LLM Agents

cs.NE ·2026-10-18 ·Kargi Chauhan, Sadiba Nusrat Nur, Pratibha Revankar, Kyle Montgomery, Chenguang Wang
publishedbenchmark
Abstract · Baylearn2026 paper accepted for Oral Presentation

Agents that run for a long time have a memory problem, and it is not the one most systems optimize for. Storing every trace and retrieving the nearest neighbors works early and quietly stops working later: the store fills up with failures the agent then repeats, and with facts that were true fifty turns ago. RSPM (Recursive Self-Pruning Memory) borrows the fix biology already found. At intervals the agent sleeps. It clusters its recent failures, distills each cluster into a short rule, and deletes the raw episodes it just learned from. Rules decay unless something reinforces them, and rules that recur across many contexts get promoted into general ones. We tested this on three settings with little in common: a difficulty-calibrated SWE-bench simulation, temporal consistency across 12 conversational benchmarks, and long-horizon QA on LoCoMo-10. All three improve, by 9.1, 11.9, and 7.6 points respectively. More interestingly, the two studies that instrument the store itself converge on nearly the same operating point without being tuned toward it: roughly 93% fewer retry loops and roughly 30% fewer stored entries. That hints that knowing what to forget is a general property of agent memory rather than a benchmark-specific trick.

0.26

Operator-Mismatch Contrastive Learning for World-Relative Trajectory Coherence

cs.LG ·2026-10-18 ·One Octadion, Firmansyah Sundana, Novanto Yudistira
benchmarktheory
Abstract · 19 pages, 6 figures, 11 tables

World models are increasingly judged by how well they rank plausible futures, yet plausibility alone is the wrong target. The same scene can evolve under many latent rules, and a future that is smooth and physically valid under one rule can be incoherent under the rule actually in force. We formalize the resulting failure, operator-blindness, and show that it is systematically hidden by standard evaluation: a scorer can attain near-perfect generic ranking while separating current-world from other-world futures no better than chance. We introduce Operator-Mismatch Contrastive (OMC) learning, whose semantic hard negatives are valid-but-wrong-operator trajectories, and a matching protocol that measures world-relative coherence directly. On a controlled synthetic benchmark with in- and out-of-distribution operator parameters, across two trajectory-scoring backbones, and on feature-level external validation with CoPhy ballsCF, OMC raises operator-sensitive scoring far above a chance baseline that nonetheless achieves 0.99 generic top-1. We prove that the wrong-operator negative is necessary both for the protocol to detect the failure and for the objective to fix it, and position OMC as an auxiliary coherence signal (a trajectory-coherence scorer, not a generative model, planner, or control agent) that can be attached to existing world-model representations.

0.26

Towards Human-AI Complementarity for AI Safety

cs.AI ·2026-10-18 ·Sander Lanbo Shi, Miranda Anna Christ, Eric Modesitt, Andrew Louis, Konrad Cheng +18 more
Abstract

Safely overseeing AI is increasingly challenging for the humans doing it, as they must evaluate longer and more complex outputs with every new model generation. To mitigate this challenge, we could leverage complementarity: where humans and models make different mistakes, combining their judgements could yield an oversight signal stronger than either produces alone. We study this opportunity in realistic, safety-relevant settings, contributing the largest known suite of such tasks (3,957 instances across 8 datasets and 6 domains) as well as a platform and data pipeline for measuring human, AI, and human-AI performance on matched items. We collected 35,613 judgements from 416 human raters, finding a considerable opportunity for a complementarity effect. We also found that with the assistance methods we tested, human raters corrected the AI's errors less often, even when overall human accuracy rose, and the hybridisation methods we tested added little over the better single source on average. While complementarity is an important and feasible engineering target for safe oversight, achieving it remains an open challenge for future work.

0.26

VLA-Rolling: Fewer Denoising Steps, Better Control via Adaptive Action Reuse

cs.AI cs.RO ·2026-10-18 ·Sicheng He, Yi Pan, Qiuting Yu, Jiaming Huang, Qi Lu +8 more
benchmark
Abstract

Flow-matching vision-language-action (VLA) models generate an action chunk by iteratively denoising Gaussian noise, but typically execute only a short prefix before replanning with updated observations. Under this receding-horizon paradigm, the unexecuted remainder---the action tail---is discarded, and the next chunk is regenerated from fresh noise. We show that this discarded tail retains transferable denoising progress: action-tail initialization can match or outperform a native 10-step Gaussian baseline using fewer denoising steps. However, the optimal denoising budget varies substantially across samples, indicating that tail reuse should be adapted to the updated environment. Based on this observation, we introduce VLA-Rolling, a lightweight inference framework that reuses action tails through three stages: action-tail inheritance, per-action confidence estimation, and confidence-guided refinement through renoising and denoising budget allocation, while the underlying VLA remains frozen. Across $π_{0.5}$, SmolVLA, X-VLA, and MolmoAct2 on LIBERO, MetaWorld, LIBERO-Plus, and RoboTwin 2.0, VLA-Rolling reduces action-chunk inference cost in all 16 model--benchmark pairs, achieving up to a $1.48\times$ speedup, while improving success rate in 14 pairs by up to $15.08$ percentage points. These results show that action tails are not merely discarded predictions, but reusable intermediate states that enable more efficient and accurate VLA inference. These results demonstrate that the action tail can support the generation of better actions with fewer denoising steps. Our project is available at https://chenbao123ww.github.io/VLA-Rolling/.

0.26

World Model Interpretability: From Decoding Representations to Understanding Latent Worlds

cs.LG cs.AI ·2026-10-18 ·Jize Wang, Yongcheng Jing, Xinyi Le, Cailian Chen, Xinping Guan +1 more
position
Abstract

Emerging Joint-Embedding Predictive Architectures (JEPAs) are moving beyond representation learning towards latent world models that support prediction, planning, and control. Yet their latent worlds naturally lack a human-interpretable interface. In this Perspective, we argue that as JEPA latents become internal states for prediction and action, interpretability must move beyond what they encode to how latent variables interact, evolve, and causally drive prediction. To this end, we envision a four-stage research agenda, involving Ground, Structure, Verify, and Act, to guide this shift toward actionable latent-world understanding. Within our framework, existing JEPA interpretability efforts provide early evidence for Ground, while systematic analysis of latent structure, causal verification, and intervention remains largely unexplored. Overall, this Perspective aims to call for a broader latent-world interpretability agenda that shifts attention from decoding representations toward understanding, validating, and acting on latent worlds.

0.26

CoEvo-RSI: When Harnesses Evolve, Which Models Should Follow?

cs.LG cs.CL ·2026-10-18 ·Jiaqi Su, Guibin Zhang, Jie-Jing Shao, Cong Pang, Tiankuo Yao +3 more
community pick
Abstract

Model--harness co-evolution has emerged as a prominent paradigm for recursive self-improving (RSI) agents, jointly evolving their execution scaffolds and intrinsic capabilities. We study the persistent yet often overlooked brain--body mismatch problem in this paradigm, where an evolving harness (body) demands behaviors its model (brain) has not learned or restricts capabilities it already possesses. We introduce CoEvo-RSI, a RSI framework that advances model--harness co-evolution through agent-driven, bidirectional historical inheritance. Drawing on a shared evolutionary archive and execution feedback, a meta-agent coordinates harness development from historical designs with model checkpoint selection or merging and subsequent adaptation under the selected harness, aligning inherited capabilities with evolving execution requirements to reduce brain--body mismatch. Using only supervised fine-tuning (SFT) on its own rejection-sampled trajectories, CoEvo-RSI makes Qwen3.5-9B competitive with larger models across diverse vertical applications, including multi-agent failure attribution, role-aware text-to-SQL, and smart-home instruction following. It surpasses DeepSeek-V4-Pro on TraceElephant (+14.16 agent accuracy) and GLM-5.3 on HomeBench VM (+4.90 success rate), underscoring the need for brain--body alignment and the potential of bidirectional historical inheritance to advance agent co-evolution.

0.26

Trackable Skill Library Lifecycle Management for Agentic Reinforcement Learning

cs.LG cs.CL ·2026-10-18 ·Xi Zhu, Yuanlin Duan, Changyi Yang, Kai Mei, Riyang Bao +2 more
code/data
Abstract

Skill-augmented agentic RL has emerged as an effective approach to improving LLM agents on long-horizon tasks. Despite this progress, individual skills may follow distinct trajectories as the policy evolves, with some becoming redundant after the model internalizes their capabilities, others remaining valuable as external scaffolding, and still others turning harmful. However, existing skill library systems fail to identify each skill's true contribution due to free-riding among co-retrieved skills, difficulty confounding across tasks, and the absence of skill-specific counterfactual evidence, which prevents reliable skill lifecycle management. We propose TrackRL, an agentic RL framework where skill tracking drives skill library full lifecycle management. Specifically, within each GRPO group, we turn grouped sampling into online randomized controlled experiments. Independent random injection and within-group contrasts on the same task isolate each skill's true contribution, and reliability-weighted aggregation accumulates evidence across groups. As evidence accumulates, we continuously track skill utility, which measures the skill's causal contribution as external scaffolding, and model mastery, which reflects how far the corresponding capability has been internalized. Along this line, we formulate skill lifecycle management as an evidence-gated state-transition process that governs each skill from generation and admission through protected incubation to retention, retirement, or removal, while a soft capacity bound controls library growth. Experiments show that TrackRL outperforms recent skill-augmented RL baselines in both library-assisted and skill-free evaluation, with further analyses revealing how skill value shifts during policy learning and how evidence-driven state transitions shape skill library evolution. Our code is available at https://github.com/xizhu1022/TrackRL.

0.26

Label-Space Mismatch Distorts Cross-Cohort Evaluation in Breast Cancer Molecular Subtyping

cs.LG cs.AIq-bio.QM ·2026-10-18 ·Suxing Liu, Byungwon Min, Zhijian Zheng, Jiahao Wang
Abstract

Cross-cohort evaluation of breast cancer subtype classifiers is unreliable when source and target cohorts use different annotation standards: METABRIC labels six classes, PAM50 defines five, and a claudin-low prediction cannot be correct under a five-class standard. A classifier transferred from 1,369 METABRIC patients (train n=958) to TCGA-BRCA (n=987) and GSE21653 (n=266) separates three evaluation operations routinely conflated in cross-site comparisons. (1) Label-space restriction - an evaluation-rule change, not a retraining gain - is the dominant correction (73.76%->79.74% on TCGA-BRCA; 63.16%->68.42% on GSE21653); (2) EM prior correction on the reconciled space is an optional calibration step (Brier fidelity up, macro F1 down; to 80.65% and 74.06%); and (3) the target label space is not recoverable from unlabelled predictions (EM retains prior 0.0751 on the absent class; calibrated detectors fail), so compliance must come from cohort documentation. The protocol exposes why internal validation overstates cross-cohort value: multimodal fusion gains do not transfer (no configuration beats mRNA alone, n=694; McNemar p=0.13-1.00), and the genomic panel adds nothing to pathological complete response prediction over clinical variables (DeltaAUC = +0.0045). Clinically, without label-space compliance, cross-site accuracy comparisons misattribute annotation-standard differences to model skill; Luminal A/B and claudin-low boundary errors map to distinct treatment decisions; and subtype accuracy is a weaker portable endpoint than survival discrimination (external C-index 0.577-0.584). Reporting the cohort's allowed label set and applying label-space compliance are therefore prerequisites before any cross-cohort performance claim.

0.26

CHSM8: A Chess Transformer that Follows Natural-Language Style Prompts

cs.LG cs.AIcs.CL ·2026-10-18 ·Ilia Badanin
code/data
Abstract

We present CHSM8, a 210M-parameter chess transformer that plays moves conditioned on a free-text description of a playing style, e.g. "go for the Sicilian Defense", "solid, safe, risk-free chess" or "You are Magnus Carlsen in his prime". The model is pretrained without search on 1.07B filtered Lichess games, using a factorized move representation and legal-move masking. It reaches 1837 Elo against a strength-limited Stockfish and, in 1,000-game matches, beat every other public model we tested that was likewise trained without preference tuning, rewards or search, including one with 13 times as many parameters. We then fine-tune it to follow text on a synthetic dataset: Gemma-4-31B describes the playing style of 3M game sides, a 1B Gemma-3 model distilled from the 2.4M of those descriptions that contradict nothing on the board labels the rest, and every one of 30M selected game sides gets five descriptions. Supervised fine-tuning attaches a frozen text encoder through gated cross-attention, and direct preference optimization (DPO) uses the same moves under a mismatched description as the rejected sample. After DPO, style adherence rises on every prompt we test (for example, queens traded early 53% -> 67%, pawn moves 29% -> 32% of all moves, queenside castling 97.3% -> 99.7%), and strength without a prompt stays within noise of SFT (1792 vs. 1797 Elo). Design choices at every stage were made by fixed-budget automated search loops on top of standard architectures and training recipes. We document the full pipeline and the decisions behind it, report the ablations that failed, and evaluate the models across many style prompts and kinds of games, with 1,000 games per setting. The work is fully open source and reproducible: weights, datasets with all training labels, prompts, configurations, inference code and playable demos are public and free to build on.

0.26

Dialogue Generation for Conversational Recommendation Using User Search Interactions and Knowledge Graphs

cs.AI cs.CLcs.LG ·2026-10-18 ·Minsung Hwang, Donggyu Yoon, Gowoon Hwang, Kwanghee Jang, Joyce Jiyoung Whang
published
Abstract · 6 pages, 4 figures, 2 tables, The 3rd Workshop on Evaluating and Applying Recommender Systems with Large Language Models (EARL) at The 35th …

Large-scale conversational recommendation datasets typically require manual dialogue construction or annotation. We propose SearCl, which automatically generates conversational recommendation dialogues grounded in user search and click behaviors. SearCl segments user interactions into query sessions and derives each recommendation target from content selected after a search query. A configurable probabilistic state machine controls the dialogue flow, while two LLM agents acting as the seeker and recommender generate utterances according to turn-level constraints. To support recommendation, the recommender is provided with knowledge graphs constructed from metadata about the content items used in each dialogue, capturing their attributes and relationships. Using SearCl, we generate 3,538 dialogues from content search and click interaction data, covering 1,731 unique target items. Under an LLM-as-a-judge protocol, the resulting dialogues achieve higher quality scores than dialogues from existing benchmarks. These results demonstrate that SearCl enables controllable and knowledge-grounded dataset construction without human annotations.

0.25

TrustAgent-RFF: Agentic Diagnostic Planning for Reliable RF Fingerprinting under Domain Shift

cs.LG cs.AI ·2026-10-18 ·Chuan Liu, Tianya Zhao, Yiting Wang, Junqing Zhang, Xuyu Wang
Abstract · 10 pages, 6 figures, 7 tables. Submitted to INFOCOM

Radio frequency fingerprinting (RFF) models can achieve accuracy under source domain conditions but become unreliable when the receiver, distance, collection day, or environment changes. Incorrect predictions in unseen domains may receive high softmax confidence, making confidence thresholding insufficient for trustworthy deployment. We propose TrustAgent-RFF, a prediction level reliability framework for RFF that operates with a frozen in-phase and quadrature (I/Q) classifier without target domain labels. For each device prediction, the framework acquires radio frequency evidence: counterfactual stability, prototype consistency, acquisition domain similarity, and signal quality. A source calibrated trust estimator combines these diagnostics with classifier confidence, prediction margin, and evidence availability indicators. Source derived lower tail penalties and cross module interactions reduce trust when diagnostic signals indicate weak support. A large language model (LLM) based planner adaptively selects the next diagnostic action based on evidence, trust state, action costs, and remaining budget. A deterministic source calibrated decision layer updates the prediction state as Accept, Reject, or Defer. Experiments across five RFF datasets and classifier backbones reveal an average source to target accuracy drop of 38.7% and show that many incorrect target domain predictions remain highly confident.

0.25

Integrable building blocks for symplectic neural networks: learned-type layers, fiber-preserving layers, and what a single layer can do

cs.LG math.NAmath.SG ·2026-10-18 ·Andrew Qing He
Abstract · 27 pages, 28 figures

Exactly symplectic neural networks are built almost exclusively from shear maps and gain expressivity only through depth. We ask what else a single explicit symplectic layer can be. Observing that every layer in use is the exact flow of a Hamiltonian that is a function of commuting integrals, we obtain three families of layers: a learned-type layer, in which every conjugate pair is rotated, squeezed or sheared by a state-dependent amount and the type is found by training; moving-plane shears, residual symplectic maps outside the known ridge-function class; and fiber-preserving layers, which combine an unconstrained point transformation with a fiber translation at the cost of one linear solve. We characterize residual symplectic layers $x \mapsto x + J\nabla H(x)$: in one degree of freedom they are exactly the shears along a line, while from two degrees of freedom on a strictly larger class exists. Conjugating a learned-type layer by a symplectic network yields an exact one-parameter group with an exactly conserved learned Hamiltonian. In experiments on sixteen systems with one to forty degrees of freedom, against SympNets, HénonNets and learned-plane shear networks, layers built on nonlinear integrals are far more accurate when the Hamiltonian contains functions of those integrals. On nonlinear Schrödinger lattices with six to forty sites, twist layers alternating with linear symplectic maps are over a hundred times more accurate than every shear-based network, and with an on-site potential of learned form three to four orders of magnitude more accurate. On chains of Tao oscillators with the step size as an input, stretch layers are 45 to 70 times more accurate than learned-plane shears and recover the Hamiltonian to 1%. On generic non-separable systems, fiber layers improve on learned-plane shears by at most a factor of two, and widening the networks does not help any architecture.

0.25

When Planning Problems Become Hard for LLMs: Separating Structural Complexity from Semantic Interpretability

cs.AI ·2026-10-18 ·Mayukh Debnath, Adeyinka Adedigba, Rammohan Mallipeddi
Abstract · 10 pages, 1 figure, 17 supplementary pages

Large language models (LLMs) struggle with classical planning, but benchmarks rarely distinguish structural difficulty from the difficulty of interpreting symbolic vocabulary. We examine seven STRIPS domains with BFS-optimal plan lengths. For predicting structural difficulty without solving an instance, the zero-training delete-relaxation heuristic $h_{\mathrm{add}}$ outperforms learned alternatives on held-out domains (mean Spearman $|ρ|=0.931$), including ridge regression ($0.707$), object count ($0.657$), and the episodic retrieval architecture ARC ($0.400$). ARC also fails to outperform a linear model using the same features. To isolate semantic effects, we progressively rename predicates while preserving states, transitions, and optimal plans. Valid LLM planning success drops from $67.5\%$ to $40.0\%$ between $0\%$ and $100\%$ renaming, while refusals remain at $0\%$. An LLM supplied with ARC's conditioning information predicts difficulty in context better than trained ARC (mean $|ρ|=0.826$). Hybrid routing between an LLM and a symbolic planner depends more on calibration to the target LLM's outcomes than on the structural score selected. Thus, established symbolic heuristics already estimate structural difficulty well; improving hybrid planning requires modeling LLM behavior sensitive to predicate semantics rather than developing more structural predictors.

0.25

Bridging Concepts: Multi-Agent Reasoning over Scientific Concept Graphs for High-Impact Research Prediction

cs.AI cs.IRcs.LG ·2026-10-18 ·Donggyu Yoon, Minsung Hwang, Juneha Baek, Donghyuk Shin, Joyce Jiyoung Whang
publishedbenchmark
Abstract · 5 pages, 2 figures, 4 tables, The 3rd Workshop on Knowledge-Enhanced Information Retrieval (KEIR) at The 35th ACM International Conference o…

Within a research community, new topics often emerge when previously unconnected concepts are first studied together. We construct scientific concept graphs (SCGs)--domain knowledge graphs whose nodes are fine-grained research concepts extracted from paper text by LLMs and whose edges capture their co-occurrence in papers--and study the task of forecasting high-impact concept pairs, i.e., discovering concept pairs that are currently unconnected but will have high impact when connected. To this end, we introduce new benchmark datasets constructed from CIKM, KDD, and SIGIR conference papers, with varied prediction settings, including impact forecasting for emerging concepts unseen during training. We propose MiroGraph, a Multi-agent reasoning framework with scientific concept Graphs, in which a feature designer agent designs predictive features for a given concept pair from structural, semantic, and trend perspectives, and then three LLM agents compute the actual feature values from the enclosing subgraph of the concept pair. Using these feature values, a classifier designer agent constructs a classifier that predicts the final impact labels. MiroGraph outperforms various impact forecasting and temporal link prediction methods.

0.24

Data-driven forecasting of tropical cloud forest phenology: A dynamical systems and machine learning hybrid approach

cs.LG ·2026-10-18 ·Pedro Garcia, Saul Flores
theory
Abstract · 21 pages, 4 figures

Forecasting how species and ecosystems respond to climate change is a critical challenge in ecology, particularly in hyper-diverse tropical biomes. While traditional phenological models often rely on preconceived functional forms, this study introduces a data-driven hybrid approach that integrates dynamical systems theory with machine learning to forecast flowering and fruiting cycles. Using a robust 35-year dataset from a Venezuelan tropical cloud forest, we reconstruct the system's state-space using delay vectors inspired by Takens' theorem, incorporating climate forcing variables such as temperature and precipitation. The model was trained using Gaussian Process Regression, providing both point predictors and explicit uncertainty estimates. Our results clearly suggest that the synergy between dynamical systems and machine learning effectively captures the non-linear complexities of ecological time series. Specifically, the model achieved higher predictive accuracy for flowering and fruiting. This hybrid framework offers a scalable tool for anticipating phenological mismatches and informing conservation strategies in the face of global environmental change.

0.24

CorrectRTL: Boost RTL Generation Quality Through Automated Self-Correction with Instruction Decomposition

cs.AI ·2026-10-18 ·Jing Wang, Shang Liu, Hangan Zhou, Wenji Fang, Yuchao Wu +2 more
publishedbenchmark
Abstract · This paper is accepted by ICCAD'26

Large language models (LLMs) have shown strong promise in specification-to-RTL generation. Among the explorations, the self-correction paradigm has significantly advanced LLM-based RTL generation. However, two challenges limit further progress. Existing benchmarks are saturated by advanced LLMs, leaving little room to evaluate the agentic self-correction framework. The monolithic self-correction paradigm rewrites the entire design at once, where LLMs are susceptible to error confusion and corruption of originally correct code. To address the benchmark challenge, we contribute ScaleBench, a more challenging benchmark of 200 complex RTL designs. To address the challenge in methodology, we propose CorrectRTL, an agentic framework that realizes a new decompose-then-correct paradigm, in which the LLM agent explicitly decouples scaffold construction from detailed implementation. CorrectRTL contributes three core techniques: (1)~\emph{Structure-oriented decomposition}, (2)~\emph{Snippet-based RAG}, and (3)~\emph{SVA-based self-correction}. On ScaleBench, CorrectRTL improves Pass@1 from 20.5% to 49.0% (+28.5%). On VerilogEval, CorrectRTL achieves up to 96.1% Pass@1, demonstrating consistent improvements across both simple and complex designs.

0.24

Finite Interaction Lens: Structured Precision Sharing for Vision-Language Models

cs.LG cs.CV ·2026-10-18 ·Nozomu Fujisawa
Abstract · 32 pages

Identifying which visual interventions interact in a vision-language model requires comparing their joint and separate effects. The resulting four-output difference can be small enough for rounding error to obscure it. We introduce the Finite Interaction Lens (FIL), a mixed-precision method for evaluating finite intervention families. At each selected linear projection, FIL decomposes each candidate's input into a native state, two single-intervention components, and a full-dimensional mixed residual. It projects the single-intervention components at higher precision and reuses the same numerical contributions in the joint candidates, while retaining every mixed residual and nonlinear trajectory. We derive a local cancellation identity and a conditional bound on error propagation through a smooth nonlinearity. In one MLP of Qwen3-VL-4B-Instruct, FIL reduces final interaction MSE by 94.4% on 16 synthetic families and by 99.85% and 99.35% in two separate four-image COCO studies, relative to a control that has the same nominal BF16-product budget and higher-precision single interventions but projects each joint candidate independently. One image accounts for most of the absolute gain in each COCO study. Higher-accuracy multi-product BF16 arithmetic on all candidates gives lower aggregate error at similar total query times. Single-product BF16 already recovers the reference choices in both candidate-selection tasks we tested. On these development inputs, FIL improves interaction fidelity at a fixed nominal precision budget. Its practical value depends on a task that needs the extra fidelity.

0.24

Learning Which Device to Call: Online Tool Grounding for LLM Agents in IoT

cs.AI ·2026-10-18 ·Yichen Liu, Xiaochong Jiang, Haoran Yu, Lifei Liu
Abstract

The Internet of Things (IoT) gives large language model (LLM) agents access to physical observations, but one requested capability is usually offered by several devices, and the agent's plan does not say which one to call. Existing work retrieves relevant tools and checks plan feasibility, without learning which eligible device returns a useful observation. Unlike previous works, we study online tool grounding: binding an already specified tool call to one physical endpoint under service cost and device participation constraints. We propose a framework that learns which physical device should execute each agent-issued call. It formulates selection as a contextual bandit problem and combines adaptive context partitions with hierarchical endpoint trees, so that execution feedback is reused across related requests and devices. Selection balances task success against service cost while respecting device capability, authorization, battery reserves, and availability. Gaussian reward aggregation provides an optional mechanism for protecting individual feedback values. Theoretical results establish a finite-horizon regret bound and differential privacy for released reward statistics under fixed, reward-independent assignments. In a simulated loading-dock workload, the proposed router achieves higher step success and better cost-adjusted service, at lower simulated device energy and latency, than nearest-device, advertised-quality, UCB1, and LinUCB selection.

0.23

Multifidelity Monte Carlo Estimator for High-Dimensional Quantities of Interest

cs.LG cs.CE ·2026-10-18 ·Zhibo Dai, Hamdi A. Tchelepi, Daniel M. Tartakovsky
unusual
Abstract

Multifidelity Monte Carlo (MFMC) reduces the cost of estimating the statistics of quantities of interest (QoIs) computed via expensive high-fidelity model by using cheaper low-fidelity models as control variates. We extend MFMC to vector-valued QoIs by replacing the scalar control-variate coefficients with matrices and minimizing a weighted mean-squared error. The optimal full-matrix coefficients are available in closed form, and their variance reduction is summarized by an effective correlation that takes the place of the squared Pearson correlation in the closed-form MFMC design. Restricting the coefficients to block-diagonal, diagonal, and scalar classes gives a nested hierarchy of estimators with increasing variance reduction and increasing pilot-data demand, with closed-form coefficients for compatible weightings. For a scalar QoI obtained by linearly projecting the state, correcting the full state before projecting is never worse at the population level. Experiments on a Gaussian problem, a reduced-order model of an advection--diffusion problem, and a coarse-grid two-phase flow problem show that the best class depends on the pilot size relative to the number of covariance quantities it must estimate, and that reusing the pilot samples lowers the mean-squared error.

0.23

Semi-supervised learning on singular data sets

cs.LG math.AGmath.ST ·2026-10-18 ·Guillaume Valette
unusual
Abstract

We propose to model singular data clouds by subanalytic sets, and develop an approach to semi-supervised learning in this setting. It relies on a Morrey-type embedding for Sobolev spaces relative to globally subanalytic measures with possibly singular support. As an illustration, we present a greedy method based on a Monte Carlo approximation of the Sobolev norm of the candidate solutions. We report numerical experiments on point clouds sampled from real algebraic varieties, as well as a test on a classification problem based on MNIST images downsampled into $7\times7$.

0.23

From Proposal to Execution: Modeling Interaction Semantics in Shared Autonomy

cs.AI ·2026-10-18 ·Shashank Shekhar, Laurent Jeanpierre, Abdel-Illah Mouaddib
Abstract · 13 pages, 6 tables

In many human-AI systems, candidate actions are not executed directly but are first evaluated by another decision-making entity under a potentially different objective or decision model. Standard decision-theoretic frameworks do not require this interaction to be represented explicitly, as they define transition dynamics directly over state-action pairs, allowing evaluation to be absorbed within the transition model. We introduce a semantics-aware decision framework in which a candidate action is proposed, evaluated under an interaction context, and then resolved into an executed outcome. We formalize this through a joint action language and an interpretation function that maps each joint action and context to a semantic outcome. Transition and reward dynamics are induced through these semantic outcomes, thereby introducing an explicit representational layer for interaction semantics. Semantic outcomes provide explicit units for organizing interaction principles, which can be instantiated through application-specific evaluation mechanisms and reused across distinct interaction instances. This provides a structured interface for specifying context-dependent evaluation separately from the resulting execution dynamics, without requiring changes to the underlying planning algorithm. We instantiate the framework in shared human-AI autonomy, capturing interaction patterns such as advisory and authorization, and illustrating how interaction context influences decision making. The resulting formulation remains compatible with standard Markov decision process (MDP) planning.

0.23

Concurrent two-scale topology optimization of printable bone scaffolds via rectified-flow generation and exact latent-space refinement

cs.LG physics.comp-ph ·2026-10-18 ·Reza T Batley, Sourav Saha
benchmarkunusual
Abstract

Bone scaffolds must reconcile a macroscale mechanical response with microscale geometric feasibility, yet conventional approaches to their design restrict attainable constitutive properties. This work presents a concurrent two-scale topology optimization framework; a conditional rectified flow generates periodic lattice microstructures, which are then refined in latent space by exact finite-element homogenization sensitivities. The resulting unit-cell geometries and their spatial assignments are jointly optimized under mechanical and feature-size constraints. In a distal rabbit femur benchmark, all eight flow-started refinements produce printable cells with maximum relative errors under 1\% across six prescribed engineering moduli. Constructing the implant from a single tiled cell reproduces intact whole-femur compliance within 0.1\% and regional strain energy within 0.35\% under four loading conditions. In human femoral reconstruction, multiscale concurrent optimization produces a cortical mid-shaft scaffold with porosity 63\% while satisfying the compliance constraint within 2\% under walking and stair-climbing loads. It achieves an 84\% porosity on a proximal cancellous scaffold with strain-energy deviations below 0.25\% under the same loads. Analysis of the learned representations reveals distinct architectural families in latent space. Further, the attention heads of the rectified-flow transformer specialize for spatial shifts -- learning the analogue of a finite-difference stencil -- and coordinate symmetries, with ablations showing their influence on generated topologies. These results establish a route to printable, mechanically constrained scaffold designs. The framework is general enough for extending to many other multiscale structural design applications including topology-microstructure co-design in metal additive manufacturing.

0.22

CVTA: Cross-Variable Temporal Attention for Multivariate Irregular Time Series Prediction

cs.LG ·2026-10-18 ·Ankitkumar Joshi, Milos Hauskrecht
top venue
Abstract · Accepted at the Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS 2026), Main Track

Irregularly sampled multivariate time series arise in many real-world domains such as healthcare, climate science, and large-scale monitoring systems. Characterized by asynchronous observations and heterogeneity across variables, such time series pose significant challenges for forecasting. In this work, we introduce Cross-Variable Temporal Attention (CVTA), a short-term, continuous-time forecasting framework that learns prediction functions from its compact, interpretable state summarizing recent temporal dynamics across variables. CVTA uses Variable-Conditioned Embeddings to project heterogeneous variables into variable-specific latent spaces, and applies temporal attention to capture cross-variable dependencies from this state. Furthermore, we demonstrate that CVTA can be integrated as an auxiliary module into existing state-of-the-art prediction models through a dynamic meta-decision model. Extensive experiments on multiple real-world benchmarks show that CVTA achieves strong short-term predictive accuracy, supports efficient online inference, improves existing prediction models when used as an auxiliary module, and learns attention patterns that align well with domain knowledge.

0.22

M-JEPA: Predictive Self-Supervised Learning for Molecular Graphs with Scaffold-Shift Evaluation on Tox21

cs.LG cs.AI ·2026-10-18 ·Karthik Iyer, Nasser Sabar
published
Abstract · 28 pages, 9 figures. Submitted (pre-peer-review) version; the final published version appeared in J. Chem. Inf. Model

Self-supervised molecular representation learning is increasingly used to improve transfer on small, label-limited property prediction tasks, yet commonly used contrastive objectives can be sensitive to view construction and are often evaluated with protocols that report only discrimination. We adapt Joint Embedding Predictive Architectures (JEPAs) to molecular graphs by masking chemically coherent connected subgraphs and training a context encoder to regress the masked target embedding produced by an exponential-moving-average (EMA) teacher. To evaluate predictive self-supervision under realistic distribution shift, we introduce a leakage-resistant, compute-matched three-phase protocol: (i) an 80-run objective screening on ESOL with fixed compute budgets, (ii) JEPA-only refinement of masking and augmentation choices, and (iii) per-assay Tox21 evaluation under Bemis-Murcko scaffold splits with joint discrimination-reliability analysis and attribution diagnostics. Under matched compute, M-JEPA achieves substantially lower ESOL proxy RMSE than a compute-matched InfoNCE baseline (mean RMSE 2.12 vs 3.75; mean paired dRMSE 1.61; 95% bootstrap CI [1.52, 1.72]; Wilcoxon p=2.4e-4) and typically completes runs faster (~15-25 min vs ~35-45 min on a single GPU under identical hardware), albeit with higher runtime variance. On Tox21, improvements in ROC-AUC/PR-AUC are assay-dependent and do not uniformly imply better calibration; reliability diagrams and Brier/ECE metrics reveal endpoints where fine-tuning improves ranking but leaves probabilities miscalibrated.

0.22

LLM Agents as Suppliers of Physical System Models: Calibration and Acceptance in Modelica

cs.LG cs.AIeess.SY ·2026-10-18 ·Jörg Frochte, Herbert Schmidt
Abstract

To procure a model of a device, a client hands a supplier a description, sometimes with recorded data or a target response, and accepts the model against this handover. We let LLM coding agents act as the supplier on handovers with data, building Modelica models from library components for four simulated systems and a real switching magnet, and ask what acceptance certifies. Several agents reach the noise floor, but a handover admits a set of models: the description leaves parts of the structure open, and the data narrow it only to the agreed accuracy. We therefore accept by membership in this set, not by matching one reference, and add layers above behaviour: a physical invariant exposes an energy-creating model at the noise floor. Interpretability becomes a property of the delivered model, tested on a variant, read where none exists and shaped through the handover. Most failures are fits delivered with a systematic residual on their own fitting data, where the handover left open how close is close enough. In a follow-up, stating a threshold makes the agents compare against it, and more deliveries pass. The argument applies wherever a handover underspecifies a model meant to be read, not only run.

0.22

Transformers Learn Two Competing Sorting Algorithms, and Only One Length-Generalizes

cs.LG cs.CL ·2026-10-18 ·Khashayar Gatmiry, Nathan W. Henry, Chloe H. Su, Emmanuel Abbe, Samy Bengio +3 more
Abstract

We show that two-layer transformers trained to near-zero loss on a sorting task converge to one of two mechanistically distinct solution classes depending on their initialization. One solution class consists of solutions that do not use the second attention layer to sort, but also fail to length-generalize. The other, which uses both attention layers, length-generalizes perfectly, despite only being trained to sort sequences of a fixed length. By carefully choosing the weight initialization, we demonstrate that we can steer the model toward learning one solution or the other. We identify input types that are rare at the training length but increasingly common at longer sequence lengths as a source of this disparity in length-generalization, and find that layer-wise specialization of the second class of solutions allows them to handle these rare sequences more accurately. In the second class, we discover a mechanism by which the second layer attention can use the contextual information gathered by the first layer attention to reduce its search space. On the related task of finding the minimum of a sequence, we find a similar two solution-type phenomenon, but with the opposite generalization pattern, where a different form of layer-wise specialization causes the two-layer solution to length-generalize worse than the single-attention solution.

0.22

When Empathy Misses the Goal: A Benchmark for Goal Displacement in LLM Advice

cs.AI ·2026-10-18 ·Dean Ariel, Guy Laban
top venuebenchmarkdataset
Abstract · Accepted for publication at the 40th Annual Conference on Neural Information Processing Systems (NeurIPS 2026), Evaluations and Datasets Tra…

Conversational large language models can respond with warmth and reassurance while failing to advance the user's stated goal. We conceptualize this failure as $\textbf{goal displacement}$, the substitution of persona-maintenance content for goal-relevant substance. To measure it, we introduce $\textbf{GDB}$ (Goal Displacement Benchmark), comprising 230 scenarios across nine domains and $4{,}134$ human-scored responses from 18 models. A $2 \times 2$ taxonomy crosses hard constraints with emotional distress, distinguishing failures to confront infeasible goals or combine empathy with practical guidance from reverse displacement, where unsolicited problem-solving replaces requested emotional support. A deletion-test rubric assesses whether substantive help remains after purely supportive sentences are removed. Independent raters achieve weighted $κ= 0.82$ against author scores for displacement severity. Trap displacement ranges from $0.0\%$ to $37.7\%$ and is strongly negatively correlated with Chatbot Arena Elo ($ρ= -0.706$, $n = 15$). Reverse displacement reaches $41.9\%$, including substantial rates in models with negligible trap displacement. A factorial probe supports a distinction from social sycophancy. Human-scored responses in the high-complexity, low-pressure condition exhibit $35.0\%$ sycophancy but $7.5\%$ displacement, with little association between the constructs. Matched OLMo-2 checkpoints show that mixed empathy-with-constraints displacement remains at $52.3\%$ after supervised fine-tuning and falls to $13.8\%$ after direct preference optimization. GDB provides a reproducible diagnostic of whether responses advance the operationalized user goal, supporting model comparison, post-training evaluation, and regression testing. We release the dataset, annotations, and rubric under CC-BY-4.0 and the evaluation code under MIT.

0.21

BatchParScale: Sparse Batch-Mixing for Sequence Models

cs.LG cs.AI ·2026-10-18 ·Hans Peter Lyngsøe Raaschou-jensen, M. Onurcan Kaya, Desmond Elliott
benchmark
Abstract · 19 pages, 10 figures

Recent test-time scaling methods improve model performance by increasing inference-time compute. One practical form of this is parallel scaling, where a model samples multiple independent reasoning streams and aggregates their answers using methods like voting, reranking, or verification. While effective, this approach leaves the streams uncoordinated, and a discovery made in one rollout cannot influence any other. Multi-Sequence Verifiers jointly score candidate answers after generation, but the reasoning streams remain independent during inference. ParScale takes a step toward model-internal parallel scaling by learning to aggregate multiple forward passes, but its streams still do not exchange information while processing a sequence. We introduce BatchParScale, an attention mechanism that allows parallel streams to exchange information through a small set of communication positions. We evaluate BatchParScale on a new Parallel Needle-in-a-Haystack retrieval benchmark and in pretraining, fine-tuning, and cooperative multi-agent reinforcement learning. Across these settings, BatchParScale outperforms ParScale on language-model loss and remains competitive on coordination-sensitive MARL tasks. Masked pretraining runs beat an unmasked single-sequence baseline despite local sequence corruption, and LoRA fine-tuning shows the mechanism can be added to a pretrained Transformer. At the same time, fact retrieval performance for pretrained and finetuned BatchParScale models remains unclear.

0.21

A discrete-continuous-time stochastic framework for attention-deficit/hyperactivity disorder-related symptom modeling and pharmacological intervention via methylphenidate

cs.LG math.PR ·2026-10-18 ·Oana-Maria-Catalina Crisan
Abstract · 27 pages, 2 tables, 2 algorithms, 20 figures

This paper presents an original mathematical and computational framework for modeling the efficacy of methylphenidate as an optimal control policy acting on the nonlinear dynamics of attention deficit hyperactivity disorder. Neural processes in ADHD are conceptualized as stochastic processes with trajectories that inherently converge towards stable, yet dysfunctional patterns associated with clinical symptoms. The pharmacological intervention is mathematically formalized as an external control signal that reassesses the system's drift and redirects these trajectories towards neurotypical regimes. To bridge these theoretical stochastic models with empirical clinical data, latent representations are inferred using a novel deterministic hybrid deep learning architecture, combining a one-dimensional convolutional neural network and a Vision Transformer, trained on preprocessed functional magnetic resonance imaging data.

0.21

A Simulation-Driven Digital Workflow Integrating Bayesian Optimization for Early-Stage Process Design: A Methyl Acetate Production Case Study

cs.LG physics.chem-ph ·2026-10-18 ·Trung Ngoc Trinh, Trung Thanh Nguyen, Nikolaos Tsakoumis, Roman Tschentscher
Abstract · 53 pages (including appendices), 12 figures, 4 tables

Methyl acetate production via direct carbonylation of dimethyl ether and carbon monoxide involves strong nonlinear interactions among reactor operation, recycle structure, and energy consumption, making conventional simulation-based optimization computationally demanding. In this work, a data-driven framework integrating steady-state process simulation, surrogate modeling, and Bayesian optimization was developed for process intensification. A dataset of 2730 steady-state simulations was generated using a Python-based automated process simulation workflow and a structured design of experiments across the operating domain. Surrogate models were trained to predict carbon conversion yield (C-yield) and cost of specific energy consumption (C-SEC), with neural network-based models showing the best predictive performance and robustness. These models were integrated into a multi-objective Bayesian optimization framework to identify optimal operating regions. Results indicate that high C-yield (up to ~95%) can be achieved with reduced energy cost through lower recycle flow rates and compression demand. Pareto analysis revealed a nonlinear trade-off between C-yield and C-SEC, where further improvements near the conversion limit require disproportionately higher energy input. Validation against rigorous simulations showed deviations below 3% for C-yield and 10% for C-SEC.

0.21

Design and Cross-Domain Evaluation of an Agent Execution Harness

cs.AI ·2026-10-18 ·Saurav Joshi, Manas Gupta, Ashwin Venkatesh Raman, Akshay Budhkar
Abstract · 25 pages, 11 figures, including references and appendices

Completing complex tasks with language-model agents requires coordinating planning, execution, and review so the final result addresses the user's original objective. We present an agent execution harness that keeps the original objective separate from a revisable plan and an execution record. A planner assigns scoped tasks, workers produce outputs and artifacts, and a reviewer examines results against the original request before submitting or requesting additional work. Reusable domain-specific instructions guide these roles and are refined during construction using task inputs and execution trajectories, without official grading feedback. We evaluate complete harness configurations across terminal tasks, clinical responses, application building, and scientific questions. On Terminal-Bench 2.1, the harness passes 83 of 89 instances (93.26%); on HealthBench Professional, a Fable-configured system achieves a length-adjusted rubric score of 69.63 across 525 tasks. In a 53-task paired HealthBench comparison, the Fable-configured harness achieves an aggregate score 19.89 points higher than direct completion, despite a $-1.46$-point median task-level change; mean inference cost rises from \$0.29 to \$2.13 per task. We combine these outcomes with latency, documented instruction revisions, and reviewer decisions. Evaluating outcomes alongside execution trajectories and resource use clarifies how coordination supports task completion, where it introduces regressions, and at what cost.

0.21

Where Groups Split: A Probe Prior for Nonzero Within-Group Variance in RL Post-Training

cs.LG ·2026-10-18 ·Jonah Ellison, Mateo Ruiz, Owen Pell, Tomas Duncun, Hanae Fujita +2 more
Abstract

Group relative policy optimization learns from a prompt only when the rewards in the sampled group disagree. With a binary verifier, a group that is entirely correct or entirely incorrect has within-group variance zero and a zero advantage. Uniform sampling continues to spend rollouts on prompts that are unanimous now and stay unanimous as training goes on. Pass rate is a weak forecast of that future. Two prompts at the same pass rate can move apart: one into the range where a finite group splits, and one toward a ceiling where the group becomes unanimous again. Aperture estimates the later pass rate from a short probe. A small subset of the pool is trained with the same recipe, every prompt is scored on the base model and on the probe model, and the two pass rates are plugged into the probability that a group of the training size is all wrong or split. The resulting weight is a static prior. It can be used alone or multiplied by an online scheduler, and the GRPO update is left as it is. On mathematical and logical reasoning with 1.7B and 4B models, the prior improves greedy accuracy over uniform GRPO and raises the fraction of groups with a nonzero advantage. The same prior still helps when it is combined with dynamic sampling or with selective rollouts.

0.21

Fix, Not Just Resample: Bounded Compiler-Guided Repair in Lean 4

cs.AI ·2026-10-18 ·Haotian Deng, Chris Farber, Sonal Narasimha, Samskrithi Sivakumar, Yexiang Xue
benchmarktheory
Abstract

Automated theorem provers often rely on sampling language models, but proof assistants like Lean reject each failed step with a precise error message --- information that pure sampling discards. To overcome this, we introduce a bounded, online step-level repair module. Our search samples $K$ candidate tactics at each step and repairs a failed tactic at most $R$ times. Instead of discarding failed tactics, we feed the compiler's error message to a pretrained repair model to fix individual rejected steps in place. This approach improves the solve rate over a sampling-only baseline across all tested candidate budgets ($K \in \{1,3,5,8\}$), achieving up to a $32\%$ relative gain and solving $82$ additional theorems on a $1{,}990$-theorem subset of the LeanDojo benchmark. Repair is not free: it trades extra runtime for reachable proofs. But the trade is predictable: repair pays most when sampling is constrained; its return is front-loaded, with the first repair attempt capturing $84\text{--}96\%$ of the total gain; and the gain comes almost entirely from shallow tactic-execution failures rather than structural errors such as type mismatches. This targeted repair successfully replicates on a second benchmark (Lean Workbook), nearly doubling the solve rate with zero regressions.

0.21

FHKAN: Fractional Hermite Kolmogorov-Arnold Networks for Data-Driven and Physics-Informed Learning

cs.LG ·2026-10-18 ·Arsha Farnoush, Amirhossein Sadr, Aida Pakniyat, Dara Rahmati, Saeid Gorgin +1 more
Abstract

Kolmogorov-Arnold Networks (KANs) replace fixed scalar weights by learnable univariate maps, yet most spectral KANs still rely on Chebyshev polynomials that live on a bounded interval and therefore require a saturating squashing of the pre-activation. We introduce the Fractional Hermite Kolmogorov-Arnold Network (FHKAN), a neural architecture that equips KANs with an adaptive, trainable fractional-order generalized Hermite basis. With a nonnegative shape parameter, the underlying generalized Hermite functions are orthonormal on the entire real line and uniformly bounded under a Gaussian envelope, so they act directly on unbounded pre-activations without any saturating map. We derive closed-form derivatives of the generalized, fractional, and scaled basis away from the origin and determine its order of differentiability at the origin, where the fractional power map creates a singular point. On synthetic regression, FHKAN attains the lowest test error at every network depth among nine fixed activations and the spline-based KAN. On image classification, image denoising, and sentiment analysis, the best-performing fractional Hermite block yields the highest accuracy or reconstruction quality. On physics-informed problems, FHKAN reaches mean absolute errors of order 10^-6 to 10^-5 on the Lane-Emden equation, residuals of order 10^-4 on the Burgers equation, and absolute errors below 10^-4 on a fractional delay differential equation.

0.21

Means-Ends Analysis Helps LLMs Plan in Continuous-State Domains

cs.AI cs.LG ·2026-10-18 ·Jiayi Li, Carl Kingsford
Abstract

LLM planning research has traditionally focused on discrete-state domains, where actions form a finite, enumerable set. Existing LLM planning frameworks largely assume actions are enumerable, modeling planning as a search process over candidate actions. However, in continuous-state domains, where actions form a continuum and cannot be enumerated, this assumption breaks down. We find a remedy in Newell and Simon's Means-Ends Analysis (MEA), a general problem-solving technique rooted in heuristics from human psychology. We develop Medaka, an MEA-inspired structured prompting framework that encourages recursive goal decomposition, algebraic derivation, and per-step state tracking. To evaluate native LLM planning abilities in continuous-state domains, we develop Liquidworld, a systematic testbed on liquid mixing. On Liquidworld, Medaka can improve performance across model families, suggesting that classical means-ends analysis elicits LLM-native reasoning abilities that are helpful for continuous-state planning.

0.21

Editing as Language: Teaching LLM Agents a Grammar for Native Context Editing

cs.AI ·2026-10-18 ·Yunlong Deng, Yan Li, Xing Hu, Yida Zhu, Sean Du +2 more
Abstract

A synopsis in novels folds a story into a few lines; a function in code gathers many steps with a single execution. Context editing is native to language itself. In this paper, we propose the Native Context Editing (NCE) framework, which gives language model agents a grammar for natively editing their own context. By encoding context-editing operations in a grammar of dedicated special tokens, we teach the agents to invoke edits through the language itself. We train agents on iteratively constructed trajectory segments, with each teacher edit conditioned on the evolving context. This establishes context editing as an atomic and inherent capability of LLM agents, enabling them to dynamically modify their context as they generate responses over long-horizon reasoning and interaction. Compared with context-editing methods that learn to generate tool-use code within specific external harness environments, our approach integrates context editing as an internal mechanism of the model's native generation process, enabling better generalization across diverse real-world workflows. On four long-horizon agentic benchmarks (Terminal-Bench 2.0, SWE-bench Verified, SWE-bench Multilingual, and SWE-bench Pro), NCE improves the average task success rate over the base model by 5.6 percentage points for LongCat-Flash-Lite-Sparse and by 4.5 for Qwen3.6-35B-A3B, while reducing average token usage by 36% and 15%, respectively.

0.20

Aligning Language Models with Indigenous Community Cosmovision through Constitutional AI

cs.AI cs.CLcs.CY ·2026-10-18 ·Silvia Fernández-Sabido, Jorge Reyes-Magaña, Jesús Adrián Chay Tun, Jean Carlos Buenfil, Victor González Nahuat +1 more
position
Abstract · In preparation for the Dossier "Linguistics in AI and AI in Linguistics" of the journal TextoLivre

Large language models (LLMs) are predominantly trained on corpora that reflect Western, anglophone worldviews. This asymmetry is not neutral: when a model answers questions about territory, community authority, or the relationship with nature, it does so from assumptions that may be alien to the communities using the system. This paper presents a system for aligning language models with the ethical principles of Indigenous communities in Latin America. Drawing on interviews with five Indigenous leaders and defenders, we built an operational cultural constitution grounded in an Indigenous cosmovision, with 12 principles. On this basis, we implemented a Constitutional AI (CAI) pipeline in two phases: supervised fine-tuning (SFT) and direct preference optimization (DPO). The base models were Qwen2.5-7B-Instruct and Qwen2.5-1.5B-Instruct. With only 24 preference pairs, improvements were modest but detectable in intergenerational transmission, linguistic caution, and consultation and consent. Qualitative analysis reveals that the SFT phase was responsible for the primary shift in alignment, while DPO refined specific categories. The system was deployed in a conversational interface with retrieval-augmented generation (RAG) that allows communities to feed the knowledge base, turning alignment into a participatory cycle. We argue the architecture is functional and will scale qualitatively once the corpus is expanded and computing constraints are resolved. The work constitutes a first iteration of an epistemic justice mechanism for value alignment in the Global South.

0.20

Non-Acceleration of the Heavy-Ball Method in 1D

cs.LG math.OC ·2026-10-18 ·Konstantinos Fotopoulos, Michael Helcig, Karl Deck, Jannis Alsbach
Abstract

Despite having an asymptotically optimal rate of convergence on convex quadratics, the acceleration of Polyak's heavy-ball method and a characterization of its dynamics within more general settings have remained long standing open problems. Recently, the work of Goujaud et al. [Math. Program. (2025)] made progress regarding acceleration in the context of smooth, strongly convex minimization: for a small enough inverse condition number, either the method does not accelerate on quadratics, or there exists a smooth strongly convex function such that heavy-ball cycles over the roots of unity. The constructed functions have two variables, thus leaving the question of acceleration over univariate functions open. In this work, we use a Fourier-based analysis to i) provide a proof for the non-acceleration of the heavy-ball method over the class of univariate smooth strongly convex functions. We also ii) provide a strict enlargement of the previous root-of-unity cycling region, with insights regarding the "fractional" sets of Goujaud et al. [arXiv:2502.19916], and iii) give a higher dimensional cycling instance which is not covered by root-of-unity cycling.

0.20

Who Convinced Whom? Tracing How LLMs Change Their Minds in Multi-Agent Systems

cs.MA cs.AI ·2026-10-18 ·Arnesh Batra, Aniket Khandelwal, Jashn Khemani, Krish Thukral, Arush Gumber +3 more
benchmark
Abstract

Multi-agent debate (MAD) aims to improve language-model decisions through the exchange of arguments, yet agreement after debate need not reflect independent corroboration. We investigate how peer communication influences individual decisions and collective outcomes through behavioral and internal interventions. We introduce MAD Bench, a 2,400-question benchmark spanning knowledge and reasoning, values and moral judgment, and safety and refusal, and evaluate three open-weight models using a round-matched independent-reflection control. Among answer-changing responses, a forced-choice readout before rationale generation matches the eventual answer more often in debate than in reflection (65.4% versus 22.3% on Qwen3.5-9B). Message ablations reveal model-dependent effects of peer answers and reasoning, while attributing identical text to the receiver rather than another agent substantially changes its influence. For the tested source descriptions, stated track record affects answer preferences more strongly than professional-status cues across all three models. Attention masking produces different changes in decision margins across the three checkpoints. On fully parsed, reference-labelled non-numeric questions, debate yields small estimated accuracy improvements over initial majority voting, while increasing the rate of unanimous wrong answers by 3.1-7.4 percentage points relative to reflection. Longer debates show persistent consensus and greater correlation between agents' correctness than reflection. These findings show that stronger consensus can coexist with more shared errors, motivating evaluations that assess dependence and unanimous errors alongside accuracy.

0.20

Small-scale neural simulation of Caenorhabditis elegans on Loihi-2 using NeuroML2Loihi

cs.NE ·2026-10-18 ·Danny Rosen, Ramashish Gaurav, Yang Yi
code/data
Abstract

Neural simulation is an essential methodology for studying biological nervous systems, but it remains limited by the computational costs of repeatedly updating neural and synaptic states over many timesteps. Neuromorphic hardware is a promising substrate for this class of workloads because such devices combine event-driven execution, localized memory, and spike-based communication. However, neuromorphic \textit{software} remains comparatively fragmented and difficult to integrate with the model-description standards used by the computational neuroscience community. This disconnect limits reproducible experimentation and makes direct comparison between conventional and neuromorphic simulation workflows difficult. This paper introduces ``NeuroML2Loihi'', an open-source framework for translating NeuroML model descriptions into Loihi~2-deployable Lava primitives, enabling standardized biological models to be executed on Intel's Loihi~2 neuromorphic chip. We evaluate our framework using c302-derived \textit{C. elegans} neural models across three connectome-scale configurations. Cross-platform similarity is assessed using spike-train and voltage-trace metrics, while latency, power, and energy consumption are benchmarked against the conventional reference workflow. Across the evaluated models, Loihi~2 execution enabled through NeuroML2Loihi achieves 514--1211$\times$ lower latency and a 1.8--1.9$\times$ reduction in mean wattage relative to the NEURON reference, producing substantially lower total energy consumption primarily through reduced execution time. These results demonstrate that standards-based neuromorphic neural simulation is feasible on Loihi~2 and can support meaningful comparison against established computational neuroscience workflows.

0.20

System X: An LLM Decision Model That Replies from Its Own KV Cache

cs.AI cs.CL ·2026-10-18 ·Syed Usama Khalid Bukhari, Syed Safwan Khalid, Muhammad Aamir
Abstract · 25 pages, 5 figures, 10 tables

Decision models return probabilities over supplied options, but a deployed agent may also need to reply in words. The present study introduces System X, a large language model (LLM) decision model that adds a text-reply option to the LLM-as-Jev readout. When this option wins, or confidence is low, the same model continues from the decision's key-value (KV) cache, avoiding a repeated prompt prefill. Qwen3-0.6B has been fine-tuned with LoRA on Reason-9k and Bitext reply rows. JevBench accuracy shows no statistically significant change (55.4% to 54.5%, p=0.87), while expected calibration error falls from 0.292 to 0.169. Held-out Bitext routing accuracy reaches 98.3%. Cache reuse saves 15-37% of latency on an Apple M4 Pro, with token-identical replies in the FP32 demonstration cases. Scoring options on copies of the prompt cache matches full-sequence scoring to within 1.4e-6 in probability. Adding dialogue, industry and complaint data (System X+) improves support-task results but lowers IFEval accuracy from 55.5% to 45.0%. System X v4 uses Qwen3-1.7B, rebalanced data and teacher-rewritten reply targets. Against its base LLM, JevBench accuracy rises from 55.0% to 61.0% (p=0.04), with no statistically significant IFEval change (p=0.71). A matched latency test on unadapted Qwen3-1.7B saves 17-39%. One LLM judge prefers v4 replies to an unadapted 1.7B second call (31 to 11 pairs, p=0.003); the other's preference is not significant. These reply comparisons also differ in prompts and generation limits. State conflicts and unsupported reply details remain failure modes. Further studies with several seeds, human raters and real support data are recommended.

0.20

Beyond the Teacher: Understanding and Preserving Distillation Overshoot in Agentic On-Policy Distillation

cs.LG cs.CL ·2026-10-18 ·Jike Zhong, Ming Li, Yuxiang Lai, Xi Xiao, Sheng Zhang +7 more
code/data
Abstract · Code: https://github.com/Jike338/SPOD

On-policy distillation (OPD) trains agents on their own trajectories using supervision from a stronger teacher. We uncover a counterintuitive phenomenon in multi-turn agentic tasks: a smaller, initially weaker student can substantially outperform its teacher during training, only to regress as distillation continues. We term this phenomenon distillation overshoot. Through decision- and task-level analyses, we trace overshoot to two competing effects of OPD: repairing weak student decisions and eroding successful student choices that the teacher disfavors. Early in training, repair outpaces erosion, allowing the student to combine the teacher's corrections with its own strengths and surpass the teacher. As training continues, erosion overtakes repair, causing these gains to vanish. Motivated by this finding, we propose Success-Preserving On-Policy Distillation (SPOD), which selectively masks conflicting teacher supervision on successful student choices and introduces a soft probability floor to counteract their erosion, while retaining teacher guidance elsewhere. SPOD requires only binary episode success and model probabilities already available during OPD. Across two model families and three agentic benchmarks, SPOD consistently outperforms ordinary OPD and evaluated variants, improving final success by up to 15.2 percentage points and surpassing the teacher by up to 20.0 points. Moreover, SPOD's final checkpoints outperform even ordinary OPD's best intermediate checkpoints in all six settings, demonstrating that beyond-teacher gains can be not only preserved but further extended. Code: https://github.com/Jike338/SPOD

0.20

A failure of Imagination: VLMs fail to use visual representations for text, but self-distillation can help

cs.AI cs.CLcs.CV ·2026-10-18 ·Shashwat Singh, He He, Tal Linzen
Abstract

Vision-language (VL) models are typically built by fine-tuning a language model to attend to visual features in service of a Visual Question-Answering objective. We ask whether this alignment changes how language models represent text---especially on tasks where visualization would help, like geometry questions posed in words. We distinguish weak multimodality (the model can answer questions about an input image) from strong multimodality (the model's text representations themselves become vision-aligned, improving performance on text-only tasks with latent visual structure). To test for strong multimodality, we construct a task of shape-counting problems with paired equivalent text and image specifications. This task is trivial when the image is provided, but requires visualization when specified in text as line coordinates. We compare instruction-tuned Qwen-2.5 and Mistral models against their vision-aligned counterparts on the text variant of the task. We find that the multimodal models do no better in this setting, despite strong performance on the image version of the task. Probing experiments corroborate this: vision-aligned decoder embeddings carry no more visual structure than their text-only counterparts. Finally, we provide an optimization objective that distills in-context visual embeddings into text embeddings. This alignment induces gains on the text variant of the task without explicit task supervision, even on samples more complex than seen during alignment. This suggests that VQA-style alignment leaves available cross-modal transfer untapped.

0.19

Dynamics-Aligned Representation Learning under Induced Partial Observability

cs.LG cs.AI ·2026-10-18 ·SM Mazharul Islam, Manfred Huber
Abstract

Model-based reinforcement learning (MBRL) has enabled sample-efficient learning of continuous control policies from high-dimensional observations through learned latent dynamics models. These models are typically trained with observation reconstruction objectives, which provide strong supervision but can bias representations toward visual fidelity leading to an undertrained prior and unreliable long-horizon predictions. We propose a stochastic gating mechanism that induces controlled partial observability during training by sampling a spatial binary mask from a Beta--Bernoulli random field and applying it to the raw observation. The gate probability is resampled at each time step, exposing the encoder to conditions ranging from near-total occlusion to near-full visibility. Integration of our method requires minimal architectural changes from existing dynamics models: the gated observation and the binary mask are channel-wise concatenated and passed to the encoder as a single augmented input, with no modifications to the broader training pipeline. We evaluate the proposed approach on a set of pixel-based continuous control tasks selected from the DeepMind Control Suite. Using model predictive control for evaluation, which relies solely on open-loop model rollouts, we show that stochastic gating improves planning performance and data efficiency across tasks. Further analysis demonstrates that performance is robust to the spatial scale of the mask, and a balanced observability during training outperforms configurations biased toward either sparser or denser observations.

0.19

Stepwise Reasoning Verification: Don't Generate, Classify!

cs.AI ·2026-10-18 ·Yifei Shi, Michalis Korakakis, Rami Aly, Eric Chamoun, Lukasz Szpruch +1 more
Abstract

Verifying the reasoning of large language models is increasingly important as they are applied to complex mathematical and scientific problems. Existing work focuses on large decoder models and reports mixed results for generative and discriminative verifiers, leaving unclear which paradigm is more effective and whether large models are necessary. In this work, we cast stepwise verification as a natural language inference task: the problem and partial solution prefix form the premise, while the next reasoning step serves as the hypothesis to be classified as entailed or non-entailed. This formulation enables error localisation at individual reasoning steps and direct comparison of encoder classification, decoder classification, and decoder verdict generation across model scales and annotation budgets. Across five reasoning datasets, a 0.15B encoder classifier outperforms most generative verifiers with up to 4B parameters. On identical decoder backbones, classification outperforms generation from 0.5B to 4B parameters and matches it on the strongest 7B backbone---reaching comparable performance with roughly one tenth as many annotated problems. Our findings establish classification as a parameter- and data-efficient approach to stepwise verification and show that strong verification does not require large generative decoders.

0.19

EasyResearch: Honest Regression, Forecasting and Signal Classification for Researchers Who Do Not Program

cs.LG ·2026-10-18 ·Ahmad B. A. Hassanat, Ghada A. Altarawneh
code/data
Abstract

EasyClassifier showed that researchers who do not program can obtain honest classification results when the correct evaluation is the default. Most applied studies, however, predict a number, forecast a series or classify recordings, and each task has its own way of producing scores that do not hold on new data. We present EasyResearch, open-source analysis cores for regression, time-series forecasting and signal classification, and EasyResearch Desktop, a Windows application that runs these three tasks and EasyClassifier's classification without code. All four follow the same rules: everything learned from data is learned on training data only, models keep their published defaults with fixed seeds, the selected model receives a separate final score, forecasts are never evaluated on shuffled data, and every run writes a report with a Methods paragraph and the references for every method used. We measured whether the reported scores are honest by keeping part of each dataset untouched. In regression (9 public datasets, 90 runs), the common practice overstated R^2 by +2.2 points and the best comparison score by +2.5, while EasyResearch's final score was -1.0; screening predictors on all rows before cross-validation, which EasyResearch does not offer, reported R^2 = 0.27 on pure noise against -0.44 on new data. In forecasting (15 series), shuffled cross-validation over windows of past values promised errors 19% smaller than its chosen models achieved in the untouched future, while EasyResearch's final-period score was unbiased to within 1.5%; the models EasyResearch chose had 14% lower future error, and on random walks shuffled cross-validation chose a machine-learning model in 5 of 5 series, each worse in the future than repeating the last value, which EasyResearch chose 4 times. In signal classification (4 UCR datasets) one method won every comparison and neither estimate was biased.

0.19

Enhanced Masked Diffusion Unlearning

cs.LG cs.AIcs.CL ·2026-10-18 ·Tomer Garber, Tom Tirer
Abstract

Masked diffusion language models (MDLMs) generate text by iteratively unmasking tokens with bidirectional context. As these models reach real-world deployment, machine unlearning becomes essential. This process aims to remove the influence of a designated "forget set" without retraining from scratch, so that sensitive, copyrighted, or restricted content can no longer be elicited while preserving retained capabilities. Masked Diffusion Unlearning (MDU) is the first publicly available unlearning method customized for these models: it tunes the model $p_θ$ to match a frozen prompt-masked anchor $p_{\mathrm{anchor}}$ on the forget set by minimizing the anchor-last divergence $\mathrm{KL}(p_θ\Vert p_{\mathrm{anchor}})$. We identify two failure modes of MDU. First, random answer masking can leave key forget tokens visible to the anchor even when the question is fully masked. Second, at a memorized (near-one-hot) prediction, the anchor-last logit gradient vanishes while the anchor-first divergence $\mathrm{KL}(p_{\mathrm{anchor}} \Vert p_θ)$ logit gradient does not. As a result, MDU training stalls on the tokens we most need to forget. Backed by theoretical and empirical analyses, we propose Enhanced Masked Diffusion Unlearning (EMDU): anchor-first KL divergence toward a frozen anchor whose entire answer span is masked. On TOFU and RWKU benchmarks using LLaDA-8B and Dream-7B, our EMDU matches or exceeds the forgetting of MDU and adapted AR baselines, while better preserving retain utility.

0.18

PACT: NVFP4 Data-Flow LoRA Fine-Tuning on Consumer GPUs

cs.LG ·2026-10-18 ·Lei Gao, Hossein Entezari Zarch, Chaoyi Jiang, Murali Annavaram
Abstract

Fine-tuning language models on consumer GPUs is constrained by computation and memory. Although LoRA reduces trainable parameters and optimizer state, frozen projections remain expensive, and activations retained for backward can dominate memory usage. We present PACT, a LoRA fine-tuning system that accelerates frozen projections with native NVFP4 GEMMs while reducing quantization and dequantization overhead and limiting quantization error. To reduce activation memory, PACT reuses quantized RMSNorm outputs and saves gate/up activations in NVFP4 for backward. To limit numerical error, it preserves BF16 inputs for forward nonlinear operations and calibrates frozen weights offline using GPTQ adapted to NVFP4 scales and quantized activations. Fused nonlinear kernels and cached forward and backward weight layouts reduce quantization and dequantization overhead. Evaluations across multiple models and tasks on a 16\,GB RTX~5080 show lower memory usage and faster LoRA fine-tuning with competitive accuracy, enabling model adaptation on resource-constrained consumer hardware.

0.18

Memory as Belief: Bayesian Evolution of Memory for Self-Evolving LLM Agents

cs.LG ·2026-10-18 ·Yiming Xu, Qian Yang, Zihan Chen, Yuefei Wu, Wenhan Zeng +2 more
Abstract

Large language model (LLM) agents increasingly rely on external memory to accumulate past experiences for continual adaptation. However, effectively reusing memory requires accurately assessing its value, as retrieved experiences may become unreliable when transferred across task contexts. Despite its importance, memory valuation remains largely underexplored, with existing methods relying on historical retrieval success rates that overlook the context dependency of memory usefulness and the varying reliability of estimates under different amounts of evidence. Inspired by the Bayesian brain hypothesis, we propose BEAM (Bayesian-style Evolutionary Agentic Memory), a framework that reformulates memory value evaluation as a parameterized cognitive inference process, where memory value is represented as a context-conditioned dynamic belief rather than a passive historical statistic. Specifically, BEAM infers memory value beliefs from current task contexts and continuously calibrates them using execution feedback as empirical evidence, transforming memory from passive storage into an adaptive cognitive mechanism for continual self-evolution. Extensive experiments across reasoning and code generation benchmarks demonstrate that BEAM consistently improves memory utilization and enables continual self-evolution of LLM agents.

0.18

On Reasoning Correctness Directions

cs.AI cs.LG ·2026-10-18 ·Munachiso Samuel Nwadike, Zayd M. K. Zuhri, Rifo Genadi, Abzal Nurgazy, Ali Mekky +3 more
Abstract

We present experimental evidence suggesting that information about whether a model's reasoning is on the right track can be linearly decodable without defining a causally useful direction for steering. To study this, we introduce ChainSearch, a controlled arithmetic reasoning task that extends Countdown with explicit search and backtracking. We train a 3-million-parameter model to solve this task. Its compact size allows us to study internal model correctness directions in detail using linear probes, mean activation differences, and direct optimisation, as part of a rigorous search for internal reasoning-correctness directions. We find that while the model contains statistically significant internal directions of reasoning correctness, none of the resulting directions consistently improves reasoning under causal intervention. This calls into question whether the linear representation hypothesis is sufficient for identifying causally useful directions in the context of reasoning. Our code is made available together with this work.

0.18

Decision-Target Reversal under Informative Label Missingness

cs.LG stat.ML ·2026-10-18 ·Farinborz Setoudehtazangi, Geoffrey J. McLachlan
Abstract

Informative label missingness can improve classification efficiency because the pattern of label availability may itself contain information about the discriminant structure. We show that this advantage need not persist across deployment objectives: with the data-generating model and statistical experiments held fixed, informative partial classification can be favourable under one loss and unfavourable under another. We call this phenomenon decision-target reversal. For binary Gaussian discrimination, we derive a local cost-sensitive excess-risk expansion and, in one dimension, obtain explicit critical cost ratios at which the first-order risk ordering reverses. We show that the required covariance geometry arises under finite uncertainty-dependent missingness mechanisms with a fixed marginal missing-label proportion. We then extend the analysis to multiclass finite-action problems through loss-specific active-boundary curvature, yielding a general decision-target frontier. Monte Carlo experiments illustrate the predicted reversals, and a descriptive gastrointestinal-lesion application with agreement-defined label availability illustrates sensitivity to deployment costs. The results show that favourable informative missingness is a property of the statistical experiment relative to the deployment decision, rather than of the experiment alone.

0.18

Learning without Overwriting: Structural Knowledge Fusion for Continual Learning

cs.LG cs.AI ·2026-10-18 ·Kanishk Bakshi, Amitansu Pattanaik, Kathiravan Srinivasan
Abstract · 20 pages, 7 figures, preprint

Catastrophic forgetting in sequential deep learning arises from a fixed-capacity architectural constraint: parameter updates encoding new tasks perturb weight subspaces supporting prior representations. We introduce the Knowledge Fusion Network (KFN), which mitigates the stability-plasticity trade-off in class-incremental learning through novelty-gated structural expansion rather than parameter negotiation within a shared capacity budget. KFN decouples task acquisition from knowledge consolidation: each incoming task is learned in an isolated lightweight specialist network (2.8M parameters) and integrated into the global model via a Structural Fusion Module that allocates dedicated channel capacity conditioned on a quantitative novelty score. Dual structural isolation--channel-space separation with temporally phased gradient unfreezing--protects consolidated knowledge. On ten-task Split CIFAR-100, KFN achieves 11.94% minimum task accuracy and 71.16% task-10 accuracy, whereas shared-parameter baselines degrade severely: LwF to 9.07% average accuracy (0.00% min), Replay++ to 8.91% (85.08% forgetting), and Replay to 16.21% (73.44% forgetting); KFN's forgetting (57.92%) is 15.5 pp lower than Replay. Furthermore, KFN prevents logit-scale drift, maintaining near-symmetric task performance. Across CIFAR-100 and EuroSAT satellite imagery, KFN's class-incremental retention (72.7% and 42.5%) substantially exceeds the best baselines (29.6% and 0.94%). Crucially, KFN achieves parameter-efficient linear scaling, adding 25% of the base parameter count per step and requiring 25.1% of naive backbone duplication over ten tasks (28.1M vs 112.0M). With a 37.1% peak GPU memory reduction over full retraining, KFN establishes novelty-gated expansion as a practical path beyond fixed-capacity architectures.

0.17

Assessing the AI Readiness of Construction Datasets: A Multidimensional Operational Framework

cs.AI ·2026-10-18 ·Xiaoyu Hou, Bo Xiao, Abdolmajid Erfani, Ronghua Xu, Yu Cai
Abstract

The rapid growth of artificial intelligence (AI) applications in construction has increased the demand for datasets suitable for reliable model development, evaluation, and reuse. However, construction datasets vary widely in documentation, quality, governance, and evaluation support, and consistent criteria for assessing AI readiness remain limited. This study develops a Construction Dataset AI Readiness Framework with 30 attributes across four dimensions: Metadata Completeness, Data Quality and Representativeness, Data Governance and Accessibility, and AI Utility. The framework uses attribute-specific scoring, applicability screening, and automated and manual verification. It was evaluated using 21 publicly available datasets across seven data-type groups. Overall readiness scores ranged from 0.720 to 0.964, with a mean of 0.879. Hybrid verification achieved 90.5% agreement between automated suggestions and manually verified scores. The framework supports consistent assessment of heterogeneous construction datasets and identification of readiness gaps.

0.17

Predicted Class Correction: Conformal Thresholds for Classes with No Calibration Data

cs.LG cs.CV ·2026-10-18 ·One Octadion, Novanto Yudistira, Yuta Nakashima
Abstract · 23 pages including supplementary material, 4 figures

Class-conditional conformal prediction calibrates a separate score threshold for every class and therefore requires labelled calibration examples of every class. In long-tailed deployments these are exactly what is missing: rare classes are left with few or no held-out examples, and existing methods are either undefined for them or fall back to the shared marginal threshold. We ask whether the per-class correction can instead be predicted from descriptors of the class that require no labels at all. We propose Predicted Class Correction (PCC), which regresses the offset between the per-class and the marginal conformal quantile on geometric descriptors of the classifier head, shrinks the prediction by a scalar selected on the labelled classes, and refits a single marginal offset. The descriptors depend only on the model's weights, so a class with zero calibration examples still receives its own threshold, and since only one scalar is calibrated, split conformal's marginal guarantee is retained. On ImageNet, PCC improves worst-class coverage on such classes by 0.058 [0.006, 0.110] at matched average set size, 42% of an oracle that sees the test labels. Of eight baselines on the same splits, six change the statistic by exactly zero and the best reaches 0.016 with an interval containing zero, so PCC is the only one distinguishable from no improvement; the gain persists with only 100 labelled classes out of 1000. A second finding concerns evaluation: worst-class coverage becomes unmeasurable before it becomes unachievable. Below roughly 35 test examples per class no improvement can be shown even for an oracle, so a null result there reflects the evaluation set, not the method. Restricting evaluation above this depth recovers a significant improvement on Pl@ntNet-300K, where the effect had appeared absent.

0.17

UpLoop: Upcycling Large Language Models with Gated Loop

cs.LG cs.AI ·2026-10-18 ·Tieyuan Chen, Huabin Liu, Yingtao Zhang, Shijie Li, Tengyu Pan +2 more
Abstract

Looped large language models trade parameter storage for recurrent computation, but most are trained from scratch. In practice, pretraining with looped depth is expensive, making post-training upcycling of a pretrained non-looped LLM an attractive alternative. We introduce \textbf{UpLoop}, a post-training recipe based on \emph{layer-wise} recurrence. Instead of looping a whole model or a middle block, UpLoop loops each selected layer locally, preserving the pretrained depth hierarchy and avoiding feature-space mismatch. To make loop execution useful, we add hidden-wise gates on the attention and MLP residual branches, so later loops act as progressive refinements rather than redundant rewrites. Empirically, UpLoop breaks the commonly observed saturation at two loops ($R=2$) in prior looped models. Starting from Qwen3-4B Base and applying SFT, UpLoop improves the average accuracy across 20 benchmarks by 6.17\% at $R=3$. We use $R=3$ as the main setting for the best accuracy--cost trade-off, while both native and extrapolated $R=4$ yield further gains.

0.17

Benchmark for Structured Retrieval-Augmented Generation in Perovskite Photovoltaics

cs.LG cs.AIcs.IR ·2026-10-18 ·Qiwei Han, Chi Zhou, Gaoxiang Yuan
benchmark
Abstract · 8 pages, 3 figures, 2 tables

While retrieval-augmented language models increasingly serve as literature assistants in the physical sciences, existing benchmarks evaluate them primarily on qualitative text overlap or through model-based scoring rather than against verified experimental ground truth. In experimental materials disciplines such as perovskite photovoltaics, however, the most critical inquiries are inherently quantitative and aggregate measurements across dozens or hundreds of papers. Here, we present a quantitative question-answering benchmark where all 755 ground-truth answers are computed directly from 45,441 measured device records (40,889 human-curated records through 2021 and 4,552 records extracted from 2022-2026 literature). Benchmarking six retrieval architectures over 158,274 passages from 25,662 papers using an open 14B model on a single 24 GB GPU, standard passage retrieval tops out at 16-23% accuracy (compared to 1.6-6.5% zero-shot). In contrast, equipping the language model with read-only SQL execution over the structured device table elevates accuracy to 51.5% on historical and 50.8% on prospective literature -- a 33.01.8 percentage-point advantage over hybrid text retrieval. Aggregate queries spanning extrema, distributions, and filtered record counts prove virtually unsolvable via text retrieval alone, demonstrating that systematic table extraction is an operational prerequisite for literature-scale quantitative inquiry.

0.17

EG-GRPO: Evidence-Guided Step-level Credit Assignment for Deep Search Agents

cs.LG cs.AI ·2026-10-18 ·Sijia Luo, Mengjie Gao, Yuxiao He, Mengshu Sun, Cuiping Li +2 more
Abstract

LLM-based search agents demonstrate strong capabilities in long-horizon information seeking and reasoning, typically trained via reinforcement learning. However, existing approaches predominantly rely on trajectory-level outcome rewards, providing limited supervision for complex search and reasoning processes. Although step-level credit assignment offers a promising solution, obtaining reliable and task-grounded supervision for open-ended search remains challenging. In this paper, we propose Evidence-Guided Group Relative Policy Optimization (EG-GRPO), a framework that derives step-level supervision from the evidence annotations available in synthetic deep search data. EG-GRPO identifies the first supporting citation for each required evidence claim and uses the newly acquired evidence to construct step-level process rewards. By representing the accumulated evidence as an evolving knowledge state, EG-GRPO further performs state-conditioned relative advantage estimation and combines the resulting step-level signal with the standard trajectory-level outcome advantage. Experiments on five challenging deep search benchmarks show that EG-GRPO generally outperforms standard GRPO and existing reward-enhanced and step-level RL baselines across different model scales. These results demonstrate that task-grounded evidence annotations can provide effective process supervision for credit assignment in deep search RL.

0.17

Synergy-Aware Residual Flow Matching for Single-Cell Perturbation Prediction

cs.LG cs.AI ·2026-10-18 ·Jingrui Wu, Wangyun Dan
Abstract

Predicting single-cell responses to genetic and chemical perturbations is important for understanding regulatory mechanisms and prioritizing therapeutic interventions. Existing methods have made progress by modeling perturbation-conditioned expression shifts with neural networks, graph priors, foundation-model representations, or distributional generative models. However, combinatorial perturbation prediction remains challenging because much of the response can be explained by additive single-perturbation effects, while the biologically informative signal often lies in a smaller non-additive residual induced by perturbation interactions. Directly generating the full post-perturbation expression profile may therefore allocate substantial capacity to the dominant additive component and underemphasize synergy-specific structure. To address this issue, we propose \textbf{SynGeneFlow}, a synergy-aware flow matching framework that anchors the source distribution at an additive prior and learns a residual flow toward the true perturbed state. SynGeneFlow further incorporates perturbation-conditioned gene interaction modeling, where a static gene-gene mask provides biological structure and a perturbation-dependent deformation bias dynamically modulates gene-gene coupling. To preserve population-level response structure, we align both endpoint distributions and biological residual distributions with MMD-based regularization. We evaluate SynGeneFlow on genetic and drug perturbation benchmarks under observed-single, held-out, and cross-modality settings. The results demonstrate that residualizing flow matching around additive perturbation effects provides a biologically grounded and robust framework for single-cell perturbation response prediction.

0.17

Improving the Expressive Power of Hybrid Sequence Models via Multi-Hop Retrieval

cs.LG ·2026-10-18 ·Yam Eitan, Gilad Yehudai, Haggai Maron
theory
Abstract

Hybrid architectures offer a promising approach to efficient language modeling by combining a small number of full-attention layers with a larger number of efficient sequence layers, such as sparse attention and recurrent layers. While full-attention layers maintain a KV cache that grows linearly with sequence length, efficient layers typically rely on separate, fixed-size states rather than accessing this richer existing memory. Recent work has begun exploring ways to allow efficient layers access to the KV cache, but the theoretical benefits of such access, and the extent to which existing approaches can realize them remain poorly understood. We investigate these questions by analyzing the expressive power of individual hybrid blocks on function composition, a well-studied task in sequence-model expressivity associated with multi-hop reasoning. We first establish that both conventional hybrid blocks and those with top-$κ$ KV-cache access require $Ω(\sqrt{n})$ capacity to compute even two-function composition over a domain of size $n$. To overcome this limitation, we introduce the \emph{Multi-hop Retrieval Attention} (MuRA) hybrid block, a simple architecture that enables multi-hop retrieval from a shared KV cache while preserving asymptotic computational and memory efficiency. We prove that a single fixed size $m$-hop MuRA block can compute $m$-function composition, whereas computing $(m+1)$-function composition with the same retrieval depth requires polynomial model size, establishing a strict expressivity hierarchy in retrieval depth. Synthetic experiments support this hierarchy, showing that increasing retrieval depth enables learning progressively deeper compositions, substantially outperforming existing hybrid architectures. Together, our results identify multi-hop retrieval as a promising mechanism for improving the expressive power of efficient hybrid architectures.

0.17

Follow the Mixed Groups: Ridge-Tracking Gaussian Sampling for Zero-Variance Web Rollouts

cs.LG ·2026-10-18 ·Tomas Duncun, Jonah Ellison, Owen Pell, Mateo Ruiz, Hanae Fujita +2 more
Abstract

Group relative policy optimization learns from a question only when the sampled responses disagree. A group that is entirely correct or entirely incorrect has zero reward variance and a zero advantage. On the Web, that group has already spent its searches and page fetches. Gaussian curricula aim the next batch at pass rates where a finite group can still mix. A Gaussian on the latest empirical rate treats every all-fail history as the same point, and a center stepped by the average signed change of those rates is pushed by prompts already stuck at 0 or 1. RIDGE keeps the Gaussian curriculum and changes the coordinate, the center, and the width. Each prompt carries a Beta posterior. The sampling weight is the marginal of a Gaussian target under a Gaussian approximation of that posterior, divided by a prediction of tool-call cost. The center and the width are an exponential smooth of the posterior means of groups that just came back mixed. All-correct and all-incorrect groups refresh the posterior and do not move the center. The GRPO update itself is unchanged. On two math backbones and a retrieval-augmented question answerer, RIDGE raises average accuracy by 3.4 and 2.9 points over GRPO and average exact match by 3.7 points, at 29.6\% fewer math rollouts and 32.7\% fewer tool calls than dynamic sampling.

0.16

Bud ContextCompact 2.0: Context Compaction for Agent Success and Accuracy with Small Language Models and Constrained Context Windows

cs.AI ·2026-10-18 ·Jithin VG, Ditto PS
position
Abstract

Agents that work for many turns outgrow the context window of the model that serves them, and on small language models (SLMs) and constrained inference systems that window is small. Context compaction replaces older turns with a summary so a session can continue, but it is lossy, costs a model call, and can help or hurt depending on when it runs. We describe Bud ContextCompact 2.0, the compaction built into the Bud agent runtime, and evaluate it in two settings. On recall questions over recorded agent sessions, compaction to a 16K budget raised a 1.2B-parameter model's accuracy to 0.335, from 0.247 with truncation and 0.176 with its full 32K history (p < 0.001), and raised a 9B model from 0.515 (truncation) to 0.577 (p = 0.04). On 100 SWE-bench Verified tasks run end to end by a 27B reasoning model on a 49K-token window (556 graded sessions), compaction raised the resolve rate on long sessions from 41% to 53% (+12.3 points, 95% CI +1.6 to +23.0, McNemar p = 0.043), left short sessions unchanged, and used 13% fewer tokens per solved task at equal total cost. The main failure without compaction was not the history overflowing but a reasoning turn exhausting the room the history left. The gain has limits: it vanished under a 45-minute session budget, because each summary took a median of 161 s, and in a threshold study it appeared only when compaction triggered early enough to leave room. An 8B model submitted after about ten turns at 5-8K tokens and never needed compaction. We argue that the accuracy results transfer across hardware, including CPU inference, for an equivalent model and window, while the time cost scales with the hardware and must be budgeted for.

0.16

DROD: Dual-Ratio Ordered Dropout for Federated Continual Learning on Heterogeneous Devices

cs.LG cs.DC ·2026-10-18 ·Berk Sezer, Özgür Soysal
Abstract · 18 pages

Federated learning for capacity-heterogeneous clients and federated continual learning are each well studied on their own, but their intersection -- clients holding differently sized models while every client's data stream also drifts over time -- remains underexplored. We study this setting directly, training a ResNet-18 fleet from scratch on Split-CIFAR-100 (10 tasks, class-incremental), and introduce DROD (Dual-Ratio Ordered Dropout), which builds on FjORD's Ordered Dropout by using two independent reduction axes to extract sub-models: a shortcut ratio and a branch ratio. On a shared parameter/FLOP budget ladder, DROD achieves the highest mean accuracy among the heterogeneous-model federated learning strategies we compare. Its advantage over the two strongest baseline geometries persists across every replay-buffer size we test and under ER, DER++, and BiC; under EWC, all three geometries collapse alike.

0.16

Symbolic Constraint Agents: Mining Formal Tool-Call Constraints to Prevent Agent Errors

cs.AI ·2026-10-18 ·Jinu Lee, Shraddha Barke
Abstract

Symbolic constraints (programmatic, deterministic checks for tool calls) enable deterministic verification of LLM agents' tool calls, but existing approaches primarily enforce predefined safety policies rather than tool dependencies in complex tool environments. We explore whether symbolic constraints can be mined from environments and observations beyond predefined policies, and whether they can improve end-to-end task performance. We present Symbolic Constraint Agents (SCA), a framework that mines symbolic constraints from API documentation, refines them against execution traces, and checks the agent's tool calls with an SMT solver at runtime. A violation triggers a natural-language warning or, when the valid replacement is uniquely determined, an automatic repair. On AppWorld, refinement raises error-detection F1 from 37.0 to 58.3 on GPT-5.5 traces and from 39.3 to 61.9 on Qwen3.5-27B traces. At runtime, repair raises GPT-5.4-mini's task success by 5.4 points under a 40-call budget and 9.8 points when environment errors are hidden, while removing 59--87\% of failed calls across models, and transfers to $τ^2$-bench with gains of up to 5.6 points. These findings show that automatically acquired symbolic constraints can improve tool-calling reliability and task completion, particularly when failures are costly or difficult to diagnose.

0.16

CasePlay: Self-Play Reinforcement Learning from Case Reports for Medical Reasoning

cs.LG cs.AIcs.CL ·2026-10-18 ·Hao Wang, Zihan Wang, Kai Wu, Jian Yao, Yiwen Ye +4 more
Abstract

Reinforcement learning (RL) is a promising way to improve the clinical reasoning abilities of large language models (LLMs), but current pipelines often rely on expert-labeled questions or costly synthetic-data generation. Meanwhile, the medical literature offers a wealth of high-quality unlabeled documents, including clinical case reports that trace the full clinical course from initial presentation to final treatment decisions. We study whether document-grounded self-play can convert these reports into effective RL training signals. We identify two challenges that limit existing self-play methods in this setting: knowledge-point fixation}, where the proposer repeatedly asks about a narrow set of salient clues in a source report, and difficulty-quality mismatch, where appropriately difficult questions may still be poorly formed or shallow, limiting their value for reasoner training. To address these challenges, we propose CasePlay, a self-play framework with two key components: (a) a Knowledge-Conditioned Proposer, which anchors question generation to case-specific knowledge points so that questions draw on a broader range of evidence from the source report; and (b) a Rubric-Judge Reasoner, which augments answer-accuracy feedback with rubric-based quality feedback on clue integration, reasoning depth, and option quality, guiding the proposer toward reasoning-intensive questions that better support reasoner training. Extensive experiments across seven medically related benchmarks demonstrate that CasePlay outperforms existing self-play baselines and improves the SFT warm-start model by 3.1 points on average, highlighted by a 6.3-point gain on LiveClin. This approach offers an effective path toward continually improving medical reasoning ability through self-play RL.

0.16

Reinforcement learning on the EPRI AI-ccelerating Unit Commitment benchmark: diagnosing over-commitment and guiding MILP with learned policies

cs.LG stat.ML ·2026-10-18 ·Maharani Devira Pramita, Aidan O'Sullivan
benchmark
Abstract

Reinforcement learning (RL) is increasingly proposed as a fast solver for unit commitment (UC), but standalone agents remain far more expensive than mixed-integer linear programming (MILP). In the first study of RL on the 51-generator, 72-hour benchmark of the Electric Power Research Institute (EPRI) AI-ccelerating Unit Commitment competition, we trace this gap to the agents' own exploration, close most of it, and show that the resulting policy can accelerate an exact solver. In every configuration tested, Proximal Policy Optimisation (PPO) and an action-branching Deep Q-Network (DQN) trained from scratch committed about twice the units of a thermal-only MILP optimum, at 39-64% higher cost, whereas a merit-order rule came within 1.6-7% (10-24% when unserved energy is priced). When each unit's decision is sampled independently, spare units insure against random de-commitments. Consistent with this, injecting exploration noise during fine-tuning enlarged a lean policy's fleet, a reward without shaping terms did not remove the over-commitment, and under their training noise the agents outperformed the rule subjected to comparable noise. Imitating the MILP solutions and fine-tuning conservatively brought PPO within 2.6% of the optimum, and these policies had the lowest penalised cost of any evaluated (about 5%). Fixing the fine-tuned policy's confident decisions made the MILP 3.8-8.5 times faster within 0.10-0.25% of the best solution, with a smaller worst case than an imitation-only policy fixing as many variables (1.1% against 8.9%). Because dispatch method and benchmark scope alone shift the reported optimality ratio by up to 3.6% and 9%, we propose a reporting protocol.

0.16

FoundPDE: A generative foundation model for data-driven symbolic discovery of partial differential equations

cs.AI physics.comp-phcs.LG ·2026-10-18 ·Langchen Liu, Junlin Luo, Min Zhu, Zhuoran Yang, Lu Lu
benchmark
Abstract · 42 pages, 13 figures, 12 tables

Discovering governing partial differential equations (PDEs) from observed data is a central inverse problem in computational science and engineering: the goal is not only to predict trajectories, but to identify interpretable symbolic laws that can be analyzed and reused. Existing PDE discovery methods often depend on hand-designed libraries, accurate numerical derivatives, or per-equation symbolic search, making them fragile under sparse observations, noisy data, broad equation families, and unseen symbolic compositions. Here, we introduce FoundPDE, a pretrained data-to-symbol foundation model for PDE discovery. FoundPDE maps discretized spatio-temporal trajectories to symbolic equations with a transformer, uses Fourier and first-order spatial-derivative augmentation to expose complementary trajectory information, refines numerical coefficients, and supports downstream extension with improved generalization through few-shot fine-tuning and hierarchical discovery. On a clean in-domain benchmark spanning 20 PDE families, our model reaches 97.50% type recognition. Under coefficient shifts beyond the training support, type recognition remains 90.53%; under multiplicative Gaussian noise on the model input, it is 94.43% at a 2% noise level but falls to 52.09% at 10%. For unseen composite equations, few-shot fine-tuning reaches over 90% exact recovery on four held-out families, and we further develop a hierarchical discovery algorithm that recovers two-term composites without fine-tuning FoundPDE. By amortizing symbolic discovery across equation families, FoundPDE shifts PDE discovery from per-dataset search toward a reusable pretrained model that extends to new physics with limited additional data, a step toward foundation models that uncover interpretable governing laws directly from scientific data.

0.15

NeuroAct: Advancing EEG Analysis and Decoding via Dual-Loop Autonomous Research

cs.LG cs.AIeess.SP ·2026-10-18 ·Wei Xiong, Jiangtong Li, Han Zhao, Linxiao Yang, Jie Li +1 more
Abstract · 30 pages, 8 figures

EEG analysis and decoding underpin brain-computer interfaces, cognitive assessment, and sleep monitoring. Fixed pipelines and reward-driven search do not explicitly determine how signal structure, representations, predictor behavior, and training dynamics contribute to performance or generalization gaps, leaving intervention choice underdetermined. We introduce NeuroAct, an autonomous research agent framework for EEG analysis and decoding. NeuroAct implements dual-loop autonomous research across EEG foundation models, task-specific predictors, and model combinations. In an inner loop, research, development, and review roles use evidence on signal structure, representations, predictor behavior, and training dynamics to test hypotheses, select analyses or interventions, and refine candidates using performance, stability, and cost feedback. An outer loop distills outcomes into conditional, testable recommendations with task and model conditions, then validates and revises them on subsequent tasks.We evaluate four settings: (A) EEG General Analysis, (B) Cross-task Model Optimization, (C) Single-task Open Research, and (D) Collaborative Prediction. In Setting A, NeuroAct's 80.88 macro-average exceeds BrainAgent's 63.13 by 17.75 points. In Setting B, NeuroAct raises EEGPT's mean balanced accuracy (BA) by 7.34 points and HBN Pearson's (r) from 0.6894 to 0.7362; in Setting C, it improves ADFTD BA by 16.33 points; and in Setting D, it raises Workload BA from 79.69% to 85.02%. Experience reuse raises BCIC-IV-2a BA from 34.66% to 50.81% while reducing LLM tokens by 64.02%, indicating recommendations improve cross-task performance. NeuroAct enables evidence-guided EEG research improved by cross-task experience.

0.15

Memory, Know Your PLACE: Governing Lifecycle State and Query-Time Eligibility in Persistent-Memory Agents

cs.AI cs.CL ·2026-10-18 ·Chen-Fang Hu, Hong-Han Shuai, Woan-Shiuan Chien
position
Abstract

Persistent memory turns prior interactions into durable agent state, but existing systems provide limited control over whether that state remains valid and whether it should influence a later query. We argue that reliable memory requires governance over both state formation and query-time influence. We introduce PLACE, a factorized governance layer that adds write-side state certification and read-side authorization to a memory backend. A factorial evaluation on Mem0 with Qwen2.5-3B and Qwen2.5-7B isolates the contribution of each mechanism. Results show that write-side governance substantially improves lifecycle state correctness, while read-side governance eliminates measured causal leakage from ineligible memory in the state-correct setting. Their combination achieves the strongest joint performance while preserving query-relevant support. These findings show that reliable persistent memory requires explicit control over both maintained state and query-time information influence.

0.14

PRECOCIAL: Pretraining with Code Curriculums via Iterative Search for Accelerated Learning

cs.LG cs.AI ·2026-10-18 ·Edward Gu, Ali Payani, Kevin Ellis
Abstract · 22 pages, 12 figures, 21 tables

Fully synthetic pretraining data generated by symbolic programs can bootstrap downstream training on naturalistic tasks, such as language modeling, but the generating programs have so far been designed by hand. We introduce PRECOCIAL, a method for automatically optimizing programs that generate synthetic pretraining data. Starting from an initial generator program, an LLM iteratively rewrites the program, and each rewrite is scored by pretraining a randomly-initialized transformer on its data and fine-tuning it on the downstream task. Because every score requires training, PRECOCIAL keeps a single network for the entire search, continuing its pretraining for each accepted program instead of training a new network per candidate, making it significantly more efficient. On four language and code datasets, at a fixed fine-tuning budget, optimized programs reach lower perplexity than hand-designed curricula, published procedural generators, random program mutation and the program the LLM writes in the first iteration of the search, and reach the same perplexity as the hand-designed curricula with 9-33% fewer fine-tuning steps. PRECOCIAL also ends with lower perplexity than FunSearch and AlphaEvolve, which train a new network for every candidate, at the same number of evaluations and a lower cost. We also observe that an optimized program generator from one domain transfers to datasets in other domains. Additionally, optimized programs allow for better downstream performance on four different reasoning and problem-solving tasks.

0.14

Outcome-Anchored Credit Reallocation for Agentic Knowledge Graph Question Answering

cs.AI cs.CLcs.LG ·2026-10-18 ·Mingyuan Li, Yucheng Li, Xiaxia Wang, Gong Cheng
Abstract

Agentic knowledge graph question answering (KGQA) requires large language models (LLMs) to solve complex questions through multi-turn interactions with a knowledge graph (KG). Outcome-supervised reinforcement learning derives verifiable outcome rewards from final answers but uniformly assigns each rollout's outcome-level advantage to all policy-generated tokens, creating a trajectory-to-operation mismatch since KG operations within a rollout can have varying effects on the search process and final answer. To address this, we propose Outcome-Anchored Credit Reallocation (OACR), which employs a training-only KG-state outcome model to estimate outcome-reward changes between adjacent KG search states and converts them into operation credit, from which token-wise weights are derived. These multiplicative weights reallocate token-level update magnitudes while preserving the sign of the outcome-level advantage at every generated token. Across three benchmarks, OACR achieves stronger overall performance than existing KGQA methods while using fewer inference-time LLM calls and tokens.

0.14

MemPrice: Online Memory Allocation via Adaptive Pricing for Long-Horizon Agents

cs.AI ·2026-10-18 ·Yiming Zong, Shiyuan Piao, Jiashuo Jiang, Fugee Tsung
benchmark
Abstract

Long-horizon LLM agents face a fundamental allocation problem as interaction histories grow: which information to retain, how much context to devote to it, and at what granularity. Existing memory mechanisms largely focus on relevance, retrieval, or compression, leaving this allocation trade-off implicit. We introduce MemPrice, which treats active context as a shared resource and dynamically allocates it across memories and representation levels. MemPrice turns accumulated memory pressure into an adaptive price on context, encouraging detail only when its estimated retention value justifies the additional footprint. This pressure persists across memory refreshes, allowing the contents of memory to evolve while preserving long-run control over context usage. We show that irreversible compression can force linear worst-case regret against hindsight following its own memory trajectory, and introduce a same-state benchmark to evaluate allocation on MemPrice's realized memory states. MemPrice enforces the hard context budget at every round and regulates long-run occupancy through adaptive pricing, without prior knowledge of the interaction horizon. Under priced-frontier regularity, its average allocation-regret bound against this benchmark vanishes when accumulated estimation error is sublinear. We evaluate the framework under matched memory budgets on LongMemEval, LoCoMo, AppWorld, and WebShop across three foundation models, spanning both query-blind memory construction and goal-directed agent context management.

0.13

How Far Can an Agent Push? Trust-Safety Margins in Live Language-Based Economic Games

cs.MA cs.GT ·2026-10-18 ·Dima Galat, Marian-Andrei Rizoiu
benchmark
Abstract

Game theory identifies when a strategic action should be accepted, but not how long a live, language-based counterpart will continue to accept it. We entered the GLEE competition in its final week with one protected agent, Milton, and four instrument agents that ran deterministic, per-game randomized trials of policy changes; Milton received changes only through pre-specified gates or registered, rollback-guarded production rules. In negotiation, learned pricing improves adjusted percentile by 0.071 (n=21,168); in bargaining, a continuation model conditioned on the rejected offer did not improve on a recency-weighted incumbent and was retired. In persuasion, a preregistered five-dose experiment (7,391 assigned games, blinded until a fixed stop) shows that recommending every unit raises sales by 0.194 of the maximum (95% CI [0.147, 0.241]) where the receiver's slack makes obedience rational (52% of the Bayesian benchmark), lowers them where it does not (-0.067), and produces a larger early-to-late purchase decline than honest recommendations (-0.123). A rule fixed before unblinding now routes Milton's dose per cell. Milton stood 11th of 460 agents in bargaining at the fixed stop; we also report the close, where the final day's pool shift moved all five agents' bargaining ratings together while their play was unchanged, and three dispatch defects were found by checking delivered behavior against the design.

0.13

Conformal Prediction-Enabled Sequence-level Policy Optimization for Reasoning Language Models

cs.LG ·2026-10-18 ·Hui Yang, Wan Tian, Tao Ren, Zhongyi Li, Yijie Peng
Abstract

Reinforcement learning improves the reasoning capability of large language models through verifiable rewards for mathematical and coding tasks. In token-level methods such as GRPO, clipping acts on individual token ratios while the reward and advantage apply to a complete response. Recent sequence-level methods align these units but still use fixed clipping bands, whose tail probabilities change as the score distribution evolves. We propose Conformal Prediction-Enabled Sequence-level Policy Optimization (CSPO), which adaptively calibrates sequence-level clipping bounds using recent training statistics. The resulting band targets a specified marginal tail budget and adapts to policy drift; calibration conditional on response length requires additional assumptions. We bound the probability that a fresh score falls outside the band by the target miscoverage level plus explicit CDF discrepancy and drift terms. Under additional moment conditions, this bound also controls the objective and gradient error relative to the unclipped GSPO power surrogate. Experiments with the Qwen3-1.7B series show improvements over GSPO on difficult mathematical reasoning benchmarks and ALFWorld, with clipping endpoints that adapt during training.

0.13

Latent World Models Should Plan with Distances Along the Latent Manifold

cs.LG cs.RO ·2026-10-18 ·Zunzhi You, Linwei Tao, Siyu Xu, Yunke Wang, Chang Xu
benchmark
Abstract

Latent world models such as LeWM plan by searching for an action sequence whose predicted terminal latent is closest to the goal latent under an ambient norm, such as the Euclidean distance. Planning with such a cost assumes that proximity in the ambient latent space reflects proximity to the goal in the environment. Studying LeWM, we find that this assumption holds only in a small neighborhood around the goal, whereas farther away the L2 cost forms spurious low-cost basins that trap the planner. Yet neighborhood-based embeddings show that the latents retain the underlying physical state. We therefore hypothesize that the latents lie near a curved, low-dimensional manifold, and that the terminal cost should measure the distance along this latent manifold rather than through the ambient space. To estimate this manifold distance, we propose the Neighborhood Graph Cost (NGC), which computes shortest-path distances on a mutual nearest-neighbor graph constructed over training latents. NGC changes only the terminal cost and requires no retraining. This single change improves the planning success rate of LeWM across all four benchmark tasks and nearly solves TwoRoom and Reacher: even at the farthest goal offsets tested, NGC reaches 95% and 100% success, whereas the L2 cost falls to 27% and 81%. These results support the manifold hypothesis and our argument that latent world models should plan with distances along the latent manifold.

0.13

Conjugate Compilation: Self-Improving Agentic Co-Compilation for Speculative Decoding

cs.LG cs.DCcs.PL ·2026-10-18 ·Annabelle Sujun Tang, Rohan Mahapatra, Christopher Priebe, Hanyang Xu, Manya Ghobadi +2 more
Abstract

As LLM adoption surges, scaling model serving requires unlocking stranded GPU capacity alongside deploying more GPUs. Speculative decoding brings two models into the serving process: a small draft model proposes tokens that a larger target model verifies. Yet compiling these models individually leaves their shared execution outside the compiler's optimization boundary. We introduce CONJUGATE COMPILATION, which reframes speculative decoding as a cross-model compilation problem and jointly optimizes the draft and target models within their shared execution rather than treating each in isolation. Our key insight is that cross-model compilation should be informed by the measured performance of the co-located shared execution it produces: an optimization that improves a kernel in isolation need not improve the models when they execute together. To expand cross-model opportunities, we devise operand layouts that interleave blocks from both models within common tiles, co-locating GEMM and non-GEMM work while preserving their mathematics. We realize CONJUGATE COMPILATION through a self-improving agentic search system in which an LLM edits cross-model optimization plans using the compiler and GPU execution as its environment. Each proposed plan is compiled and evaluated through its co-located shared execution, whose measured score, together with kernel-level verdicts and compiler feedback, informs subsequent proposals, while Monte Carlo tree search accumulates rewards to guide exploration. Across six single-GPU speculative-decoding configurations on NVIDIA B300 and RTX 5090 GPUs, CONJUGATE COMPILATION achieves a geometric-mean per-round speedup of 1.37x over conventional execution. Interleaved kernels raise this speedup to 1.45x, reaching 1.59x in the best case. CONJUGATE COMPILATION also improves GPU utilization on average by 1.46x on RTX 5090 and 1.50x on B300.

0.12

CD-OPD: Contextual Digest for On-Policy Distillation

cs.LG cs.AI ·2026-10-18 ·Mingyuan Li, Yucheng Li, Xiaxia Wang, Gong Cheng
Abstract

Teacher takeover in on-policy distillation (OPD) can produce incorrect responses even when the student and teacher independently answer the same problem correctly, a failure we quantify using Collaboration Failure Rate (CFR). To address it, we propose Contextual Digest for On-Policy Distillation (CD-OPD). CD-OPD structures the earlier context of teacher guidance through a teacher-generated dependency graph, curating the required problem conditions and preceding reasoning as node summaries alongside the unchanged guidance in a Contextual Digest before student continuation. We further develop online meta-reweighting, which compares Digest segment gradients with a task gradient from independent student responses to improve student-only performance, favoring updates predicted to reduce task loss. Across eight benchmarks, CD-OPD improves average accuracy by 1.32--7.05 percentage points over representative existing OPD methods. It also reduces CFR by 27.4\%--56.4\% relative to existing takeover-based OPD methods.

0.12

Decentralized Online Exp-Concave Optimization

cs.LG ·2026-10-18 ·Hao Qiu, Mengxiao Zhang, Yuanyu Wan
theory
Abstract

Decentralized Online Convex Optimization (D-OCO) is a fundamental framework in which a set of local agents aims to minimize a sequence of global loss functions using only local computations and communications. While recent work has established near-optimal regret guarantees for convex and strongly convex losses, the case of exp-concave losses remains an open challenge: specifically, whether logarithmic regret is achievable. In this paper, we answer this question affirmatively by proposing Decentralized Online Newton Step (D-ONS), which adapts the idea of blocking and accelerated gossip to the Online Newton Step algorithm in a non-trivial manner. Concretely, our algorithm achieves a regret bound of $O\left(Nd(1-σ_2)^{-1/2}\,\log^2(T) \right)$ where $N$ is the number of agents, $d$ is the dimension, $T$ is the horizon, and $1-σ_2$ represents the spectral gap of the network. To our knowledge, this is the first logarithmic regret bound for decentralized online exp-concave optimization. We further establish a nearly matching lower bound of $Ω\left(Nd (1-σ_2)^{-1/2}\,\log(T/N) \right)$ showing the near optimality of our upper bound. Finally, we extend our algorithm to decentralized online linear regression in which the feasible domain is unbounded and features and labels are locally observed. By gossiping covariance and label-feature statistics, we also achieve logarithmic regret.

0.12

LightCRS: Lightweight Multi-Domain Conversational Recommendation System

cs.AI ·2026-10-18 ·He Ma, Lan You, Qiyang Wu, Chen Liu
benchmark
Abstract

Conversational recommender systems have recently adopted vision-language models to match user dialogue intents with multimodal product information. However, extending these methods to many product domains faces a dilemma: deploying a separately adapted VLM per domain scales cost linearly with the number of domains and requires routing each dialogue to the right model, while naively merging all domains into a single VLM degrades accuracy because product characteristics vary widely across categories. We propose LightCRS, a lightweight, query-conditioned framework for unified multi-domain multimodal conversational recommendation. LightCRS precomputes multimodal product representations once offline with frozen encoders. Online, a compact 5.3M-parameter module fuses this evidence for each dialogue via query-conditioned cross-attention, whose parameters are shared across all domains, and compares candidates jointly with an inter-candidate Transformer. To enable rigorous evaluation, we construct MM-R2A-25, a 25-domain benchmark pairing natural Reddit recommendation-seeking conversations with Amazon product listings. LightCRS achieves state-of-the-art average HR@1, NDCG@5, and MRR, improving HR@1 by 35.4\% over the strongest baseline while reducing per-query latency by 266 times. Joint multi-domain training itself improves HR@1 by 15.2\% over separately trained single-domain models.

0.12

Near-Optimal Last-Iterate Convergence of SGD in the Smooth Interpolation Regime

cs.LG ·2026-10-18 ·Matan Schliserman, Tomer Koren
theory
Abstract

We study the last iterate of stochastic gradient descent (SGD) with a constant stepsize on smooth convex objectives in the low-noise and interpolation regimes, where the variance $σ_\star^2$ of the stochastic gradients at the optimum is small or zero. We establish that after $T$ steps of SGD on $β$-smooth convex losses with any constant stepsize $0<η<2/β$, the last iterate attains expected excess risk $\widetilde O\bigl(\tfrac{1}{η(2-βη)^3T}+\tfracη{(2-βη)^4}σ_\star^2\bigr)$, which gives the rate $\smash{\widetilde O(1/T)}$ for every constant stepsize in the interpolation regime ($σ_\star=0$), matching the $Ω(1/T)$ lower bound up to logarithmic factors and improving upon the $O(1/\sqrt T)$ rate of Attia et al. (2026) for $η=1/β$. Our analysis is based on a new technique, which compares the second moments of the stochastic gradients along the SGD trajectory across only logarithmically many time scales. For quadratic losses in the interpolation regime, an adaptation of this technique gives an $O(1/T)$ rate for every constant stepsize, extending the recent optimal rate for the randomized Kaczmarz method of Derezinski and Dong (2026) to stepsizes above the greedy stepsize with an arguably simpler proof.

0.12

Robust Direct Preference Optimization under Preference Label-Flip Attacks

cs.LG cs.AIstat.ML ·2026-10-18 ·Chenye Yang, Weiyu Xu, Lifeng Lai
Abstract

Direct Preference Optimization (DPO) is widely used for offline Reinforcement Learning from Human Feedback. However, recent works have shown that DPO can be easily manipulated by adversarial preference label flip attacks, which are specifically designed to mislead the learner toward an attacker-chosen target policy. To address this issue, we design robust DPO methods that can defend against such attacks. Building on the property that lable flips induce parameter-independent gradient shifts in log-linear DPO, we propose Trim-$K$ Robust-DPO, which removes the $K$ highest-impact pairs, and Soft-Trim-$K$ Robust-DPO, which smoothly downweights them. We establish deterministic gradient perturbation bounds uniformly over all attack sets containing at most $K$ flips, explicitly separating defense-induced bias from surviving attack effects. Under bounded-region assumptions, these bounds yield time-uniform parameter and policy deviation guarantees. We further quantify how quadratic regularization improves curvature while introducing additional bias. Experiments on Stanford Human Preferences, HH-RLHF, and UltraFeedback-Binarized validate our analysis.

0.11

Certified Sparse Attention via Perturbed Retrieval

cs.LG cs.CL ·2026-10-18 ·Runguo Li
Abstract · 58 pages (10 pages main text, 48 pages appendix), 10 figures, 35 tables

Sparse attention speeds up long-context decoding by reading few keys of the KV cache, but existing methods do not control the error of a query, and on real KV caches their errors have heavy tails. We show that perturbing the score of every key with an independent Exp(1) variable turns threshold retrieval into sampling keys with probability proportional to their attention weights. Lowering the threshold along a geometric grid gives nested samples, a Horvitz--Thompson estimate of the output at every level, and confidence bounds that hold simultaneously over all levels, so each head stops at the first level whose certificate meets the target error $\varepsilon$. With fresh perturbations at every decoding step, each certificate fails with probability at most $δ$ during generation. At a suitable threshold, the estimator reads $O(\varepsilon^{-2}\log(1/δ))$ values regardless of the context length, and no adaptive method reads fewer, up to constants. A 4-bit scan with deterministic bounds, shared across a grouped-query-attention group, generates the candidates exactly but reads every key. On three LLMs and ten settings at 16K--128K tokens, no certificate is violated, and at equal keys read our 90th-percentile error is 4.6--7.9$\times$ lower than that of the best of 15 baseline families. At equal bytes, the best baselines have 0.6--1.6 times our 90th-percentile error and at least 4.9 times our largest. End to end, our method stays within one point of dense attention on tasks and within 0.1\% in PG-19 perplexity, and the strongest baselines at equal bytes score as well. On CPU, decoding steps at 32K--128K tokens run 1.1--1.4$\times$ faster than dense attention on the same 16-bit cache.

0.11

MOSAIC: Multi-Objective Search for Allocating Inference Cache Across LLM Decoder Layers

cs.LG cs.AI ·2026-10-18 ·Sravanth Kodavanti, Manjunath Arveti
Abstract

The Key-Value (KV) cache is the main memory bottleneck of long-context LLM inference. Token eviction methods such as SnapKV, H2O, AdaKV and L2Norm shrink it, but they give every transformer layer the same budget, although layers differ in which tokens they need. We present MOSAIC, a multi-objective neural architecture search that learns per-layer KV-cache budgets for any eviction method. One search maps the trade-off between average budget and task score; its best allocation transfers to every target budget without further search, and an optional budget-constrained search refines it. Across four eviction methods on Llama-3-8B-Instruct and Mistral-7B-Instruct-v0.2, MOSAIC beats uniform allocation in 146 of 154 (task, method, budget) settings, by up to +48 points on RULER and +11.9 on LongBench (LCC). On both models, it reaches the score of uniform allocation with up to 4x less KV cache (75% smaller). Against EvolKV, MOSAIC explores a ~10^40 x larger search space with 4x fewer evaluations and 18% fewer GPU-hours per budget, and scores higher in 5 of 6 settings. The learned allocations consistently favour a small set of middle layers.

0.10

Beyond Policy Discrepancy: Capability-Aware Multi-Teacher On-Policy Distillation

cs.LG cs.AI ·2026-10-18 ·Xuanfeng Yang, Le Gai, Jiaming Wang, Xuan Zhang
Abstract

Multi-teacher on-policy distillation (MOPD) integrates domain specialists into a shared student, but determining how strongly each teacher should supervise remains challenging. Existing discrepancy-based allocation can assign strong supervision even when the student already matches or exceeds the teacher's task performance, since behavioral mismatch does not necessarily imply a remaining capability deficit. We propose Capability-Aware Multi-Teacher On-Policy Distillation (CA-MOPD), which combines specialist merging with capability-guided allocation to account for what the student has already learned. CA-MOPD first merges reinforcement-learning specialists into a strong student initialization, then uses verifier-estimated teacher--student capability gaps to adjust domain-level distillation strength. Supervision decreases as capability gaps close and increases when estimated student capability regresses, while dense token-level guidance remains provided by the teachers. Across six benchmarks spanning mathematics, code generation, and instruction following, CA-MOPD achieves an average score of 32.22, outperforming naive MOPD, Open-MOPD, and the merged initialization by 0.86, 0.83, and 0.74 points, respectively. Controlled comparisons further support the benefits of merged initialization and capability-gap guidance under matched training budgets.

0.10

Beyond the record: evaluating probabilistic AI weather forecasts deep in the tail

cs.LG ·2026-10-18 ·Bastiaan Van Velthoven, Cas Decancq, Diego G. Miralles, Stijn Luca, Thomas Mortier
Abstract

Can probabilistic artificial intelligence (AI) weather forecasts retain skill beyond their training domain? We compare four state-of-the-art AI models (ArchesWeatherGen, U-Cast, WeatherNext2 and GenCast-mini) with the physics-based IFS-ENS over land in 2022, at 1.5° resolution and lead times of 2--10 days. Monthly ERA5 records from 1979--2021 define common historical boundaries, and we characterize new records set in 2022 by their depth, i.e., the exceedance margin relative to local historical variability. Two proper scores, record-threshold weighted CRPS and Brier score, assess the magnitude and probability of record exceedances across all eligible cases, retaining both missed extremes and false alarms. Verified against ERA5, AI models outperform IFS-ENS for heat and cold records at almost every lead and depth. For heat, the advantage strengthens with depth, partly because IFS-ENS over-forecasts deep records, and largely persists for deeper records under HRES-fc0 verification. For wind, gains erode for deeper records and longer leads. Better tail scores can nevertheless coexist with aggregate probability deficits or surpluses, and some improvements arise chiefly on non-exceedances. Increasing underprediction on progressively more extreme observed subsets can also arise through outcome selection, even for an ideal conditional-mean forecast. These findings establish that native probabilistic AI ensembles retain, and for heat increase, their advantage over IFS-ENS as thresholds move beyond local records, although all advantages narrow with lead time. They caution against generalizing the extrapolation failure reported for deterministic AI forecasts to probabilistic systems, and motivate common verification conventions that separate comparative tail skill, performance on realized extremes, and probabilistic calibration.

0.09

Industrial Agentic AI: LLM-Driven New-Generation Agents for Smart Manufacturing

cs.AI ·2026-10-18 ·Lei Ren, Haiteng Wang, Shixiang Li, Jing Zhang, Yuqing Wang +5 more
survey
Abstract

Large language models (LLMs) are enabling agents to move beyond predefined workflows toward autonomous planning and execution, creating new opportunities for smart manufacturing. However, dynamic environments, complex physical constraints, and stringent reliability requirements hinder the direct deployment of general-purpose agents in smart manufacturing. This article presents a systematic overview of LLM-driven industrial agents. First, we analyze five key challenges and discuss the concept and core characteristics of industrial agents. We then establish a unified taxonomy covering application scenarios, embodiment forms, and five capability levels. Next, we propose a four-layer system architecture and detail its core modules and operational mechanisms for closed-loop task execution. Finally, we review representative applications across the industrial lifecycle and outline future research directions. This article provides a structured reference for the design, development, and application of industrial agents.

0.08

Multi-site Learning for Forecasting Day-Ahead Wind-Power Production for Onshore Wind Farms in Great Britain with Deep Neural Networks

cs.LG ·2026-10-18 ·James Kean, Aidan O'Sullivan
Abstract

Accurate day-ahead wind-power production forecasting is increasingly important for energy system operations, market activity and balancing Great Britain's grid as wind generation grows. In this study, we train multi-site deep neural networks (DNNs) using pooled historical production and reanalysis weather data from 88 onshore wind Balancing Mechanism Units (BMUs). Forecasts are produced at 09:00 UK time for the following midnight-to-midnight delivery day, matching NESO's operational day-ahead forecasting schedule. We use reanalysis weather as future covariates during training and numerical weather prediction (NWP) forecasts during inference, extending training beyond the available historical NWP archive. Our multi-site recurrent neural network outperforms site-specific RNNs and a multi-site XGBoost model, while processing the full future-weather trajectory for the forecast horizon further improves accuracy relative to hour-aligned approaches. Our best model, an RNN with a bidirectional weather encoder (BiRNN) for NWP trajectory processing, reduces MW mean absolute error and root mean squared error relative to NESO by 17.94% and 12.51%, respectively, across 638,246 matched BMU-hour observations under conditions approximating unconstrained wind farm operations. In 40 leave-one-out trials, the BiRNN achieves lower MAE than NESO at 32 excluded sites, indicating transferable forecasting ability despite reduced accuracy when sites are omitted from training. Our capacity-estimation, data-cleaning and reanalysis-based training approach provides a basis for scaling NWP-aware multi-site forecasting to additional wind farms, longer production histories, and increased forecast accuracy beyond that seen in this study.

0.08

T-MUON: Tensorizing Muon via t-SVD for Large Language Model Pre-training

cs.LG ·2026-10-18 ·Dongyang Li, Mingsong Yan, Zheng Zhang, Sui Tang
Abstract

Muon has recently emerged as a strong optimizer for large language model pretraining by exploiting matrix structure through polar orthogonalization. However, it processes each parameter matrix independently, whereas recent tensor-aware extensions of Muon jointly update groups of eligible matrix parameters through matrix unfoldings. In contrast, we propose T-MUON, a t-SVD-based generalization that groups structurally related Transformer gradients into third-order tensors and defines the corresponding Muon-style update directly through the $M$-transform. We show that the exact T-MUON update is a steepest-descent direction under the $\star_M$-spectral norm and establish a convergence guarantee. Across GPT-2-style and LLaMA-style models from 51M to 1.3B parameters trained on FineWeb and FineWeb-Edu, head-wise T-MUON with the identity transform consistently achieves lower validation perplexity than AdamW, Muon, and variants using DFT, DCT, or DST transforms. These gains persist in additional experiments with longer context lengths and larger token budgets, although the preferred tensorization mode varies across settings.

0.05

Divergence based Ranked Preference Alignment for LLMs

cs.LG stat.ML ·2026-10-18 ·Rajdeep Haldar, Yue Xing, Qifan Song
Abstract

Alignment of large language models typically relies on pairwise preference feedback, though ranked feedback can capture more comprehensive information across multiple responses within a single annotation. However, ranked-preference alignment is underexplored in practice compared to simpler pairwise approaches, potentially resulting in suboptimal behavior. Concurrently, recent research has conceptualized large language model alignment objectives as divergence estimators between aligned (preferred) and unaligned (less-preferred) response distributions, resulting in a family of objectives with an interpretable theoretical foundation and strong empirical performance. This work unifies these perspectives by introducing $f$-RPO, a $f$-divergence objective for ranked preference optimization that extends divergence-based alignment to ranked feedback and recovers pairwise preference alignment as a special case. Our theoretical analysis establishes rank-order consistency, showing that $f$-RPO preserves the underlying preference ordering at convergence. Empirical results indicate that total-variation and reverse-KL variants of the proposed objective consistently outperform ranked-alignment baselines in both ranking and generation quality across Qwen, Llama, & Gemma models ranging from 0.5B to 8B parameters.

0.05

FTFM: Factorized Tabular Foundation Models

cs.LG ·2026-10-18 ·Josif Grabocka
Abstract

Tabular foundation models (TFMs) are becoming a powerful alternative to conventional tabular learning, adapting to new datasets in context without task-specific training. Current architectures, however, face a trade-off: models that preserve cell-level representations are expensive because they repeatedly attend across rows and columns, while faster models collapse each row into a fixed embedding and lose this fine-grained state. We introduce Factorized Tabular Foundation Models (FTFM), which preserve cell representations while factorizing their computation through one contextual latent per row and one per column. Cells no longer attend to one another directly; instead, each cell is updated through the interaction of its row and column latents, reducing repeated attention while retaining feature-specific information. FTFM achieves competitive predictive performance in tabular classification on major benchmarks such as TabArena and BeyondArena.

0.04

Language, Vision & Media 133

Reranking Answers of Large Language Models with Knowledge Graphs

cs.CL cs.AI ·2026-10-18 ·Mikhail Salnikov, Hai Le, Olga Tsymboi, Ivan Lazichny, Dmitrii Iarosh +4 more
published
Abstract · 37 pages. Accepted for the Semantic Web journal (tracking no. 4075-5289); not published. https://www.semantic-web-journal.net/content/rerank…

Answering natural language questions over knowledge graph data is challenging due to the vast number of facts, which can be difficult to process and navigate. One potential solution for this issue is to use mined subgraphs related to the query, although this process still requires extracting these subgraphs. This research presents a solution for extracting subgraphs related to entity candidates from a question-and-answer set, which can be obtained by inferring a large language model by calculating the shortest paths between entities. The proposed approaches detail various features that can be extracted from the subgraphs and reranking models to select the most probable answers from a list of candidates. Experiments were conducted on Wikidata to evaluate the effectiveness of the proposed approaches. This involved enumerating all the main feature types that can be extracted from mined subgraphs and a detailed analysis of the proposed features and reranking method combinations. In addition, a public web application that provides a useful web tool for studying the graph space between question and answer entities has been developed to work with subgraphs. This includes visualization of the extracted subgraph and automatic generation of natural language text to describe it.

0.60

Beyond Factual Accuracy: A Benchmark for Evaluating Perspective Alignment in LLMs

cs.CL ·2026-10-18 ·Sunghoon Kim, Sangyeop Kim
publishedbenchmarkcode/data
Abstract

Large Language Models often generate factually correct responses that fail to maintain user-requested perspectives, blending multiple perspectives and forcing users to filter through unwanted information---a perspective control failure that factual accuracy metrics cannot detect. We introduce PerspectiveBench, a benchmark quantifying perspective alignment by constructing scenarios where multiple valid reasoning paths coexist, and propose the Perspective Alignment Score (PAS) to measure adherence to specified perspectives. Experiments reveal that state-of-the-art models successfully include target content but fail to exclude non-requested perspectives. Providing documents through RAG improves factual grounding but does not resolve perspective drift, as models prioritize comprehensive coverage over strict user constraints. Code and data are available at https://github.com/sunghoon014/PerspectiveBench.

0.52

Deep Fusion Memory: A Chunk-Level Parametric Memory for Large Language Models

cs.CL ·2026-10-18 ·Rubin Wei, Siyuan Sun, Hao Sun, Yuliang Liu, Xiang Shi +3 more
community pick
Abstract

Large language models need effective ways to acquire and use external knowledge beyond what they retain from pretraining. Continued pretraining risks catastrophic forgetting, while retrieval at inference time can add substantial overhead. External parametric memory avoids online retrieval, but constructing its training data requires extensive offline search, and output-level interpolation provides only shallow fusion. We introduce Deep Fusion Memory (DFM), a parametric memory architecture that reduces this construction workload and incorporates learned memory into intermediate hidden states. A causal memory predictor learns to generate memory from the preceding context, using retrieved chunk embeddings from the corpus as supervision. The predicted memory is shared across tokens within each chunk and fused into their hidden states through gated cross-attention. Memory prediction and fusion are trained end to end with the backbone frozen, and inference requires no explicit retrieval. Compared with existing parametric memory approaches, which often construct token-level memory targets, DFM requires $32\times$ fewer retrieval queries and $2048\times$ fewer retrieved items to construct its training data. Across six QA tasks, DFM performs better over the strongest baseline by 12.71% on Mistral-7B-v0.3 and 11.19% on Qwen3-8B-Base. Beyond QA, domain adaptation yields average gains of 6.51 and 4.38 points over Qwen3-4B-Base and Qwen3-8B-Base, respectively.

0.51

HullProbe: Measuring Interventional Part Reliance in Maritime Vessel Classifiers

cs.CV cs.LG ·2026-10-18 ·Faisal Mohammad, Mohamed Younis
publishedcode/dataunusual
Abstract · Accepted to WACV 2027 (Evaluation & Dataset Track). 20 pages including supplementary material, 12 figures, 15 tables. Code and data: https:/…

Recognizing a vessel's type from an image supports maritime surveillance and sanctions monitoring, but a classifier can reach high accuracy by leaning on the generic hull instead of the more distinctive superstructure, and is then fooled when a vessel is re-hulled or occluded. De-biasing methods claim to remove such reliance, yet the field certifies them with worst-group accuracy, a proxy that needs group labels and only infers reliance from accuracy. We measure the reliance directly. HullProbe swaps a registered anatomical part of a vessel for another instance's and reads how much that part drives the decision, using a same-class swap to control for compositing artifacts. Because a surface vessel is a hull and a superstructure arranged along one axis, its parts form a 2x3 grid whose 2^6 = 64 swaps we evaluate exhaustively, and a single Walsh transform returns each part's interventional influence and its interaction spectrum. On planted shortcuts, where the true reliance is known, the measure tracks worst-group accuracy (r = -0.94 over 45 configurations), and replacing its learned segmenter with human annotations changes the hull-share by less than the audit's resolution. On the 26 vessel types of MARVEL, the classifier places about 46% of its corrected sensitivity on the generic hull yet follows the superstructure when the two conflict, and the group-label-free de-biasers available there do not detectably reduce the hull's share. The findings hold across three architectures and a second maritime dataset, and the method extends to fine-grained vehicle recognition on Stanford Cars.

0.49

Meta Inverse Prompting for Video Generative Models

cs.CV ·2026-10-18 ·Yerin Jung, Jongheon Jeong
top venuecode/data
Abstract · Accepted to NeurIPS 2026. Project page: https://ueriniuno.github.io/mvip-page/ Code: https://github.com/ueriniuno/MVIP

Recent advances in video generative models have enabled flexible conditional generation and editing pipelines, such as text-to-video, image-to-video, and reference-guided synthesis. However, these workflows are inherently one-directional, mapping conditions to videos without providing a mechanism to invert this process. In complex media production settings, the ability to recover controllable textual conditions from a given video is highly desirable, yet remains underexplored. While inverse prompting has been studied for image generation, directly extending these approaches to video is challenging due to the need to capture temporal dynamics beyond static appearance, as well as the substantial computational cost of per-video optimization. In this work, we propose the first practical framework for efficient video inverse prompting. Specifically, we formulate inverse prompting as a meta-learning problem and introduce Meta Video Inverse Prompting (MVIP), which learns a meta-prompt optimized over temporal dimensions to enable fast and effective prompt inference. Our approach amortizes the inversion process, eliminating the need for costly iterative optimization at test time. Experimental results demonstrate that MVIP recovers higher-quality inverse prompts with significantly improved reconstruction fidelity compared to image-based inversion and iterative captioning baselines, while achieving substantial gains in efficiency. Code and video results are available at https://ueriniuno.github.io/mvip-page/.

0.49

MEMORA: A Memory-Enhanced Multimodal Committee Reranking Agent for Reasoning-Intensive Retrieval

cs.IR ·2026-10-18 ·Mohamed Mahmoud, Mostafa Farouk Senussi, Mahmoud Abdalla, Mahmoud SalahEldin Kasem, Abdelrahman Abdallah +1 more
top venuecode/data
Abstract · Accepted at EMNLP Main 2026

Reasoning-intensive retrieval is fundamentally challenged when queries demand deep semantic inference and arrive in multimodal form. Existing systems treat each query in isolation, rely on monolithic rerankers, and lack retrievers trained for cross-modal reasoning. We introduce MEMORA, a Memory-Enhanced Multimodal Orchestrated Reranking Agent built around three contributions: (1) Collaborative Factor-Committee Reranking, which decomposes queries into intent factors, extracts factor-specific evidence into a structured matrix, and synthesizes judgments via a memory-conditioned meta-verdict model; (2) an Episodic Memory Module that tracks factor outcomes across queries and dynamically adjusts priorities at inference time; and (3) a Reasoning-Aware Dense Retriever fine-tuned with chain-of-thought augmentation on 20k+ reasoning-intensive pairs. Evaluated on MM-BRIGHT across 29 domains, MEMORA achieves nDCG@10 of 41.5 (text-to-text) and 38.9 (multimodal query-to-text), outperforming the strongest agentic text-only baseline by +8.3 points and the strongest agentic multimodal baseline by +7.9 points. Code and data are available at https://github.com/HarnessLab/MEMORA.

0.48

Faithful Where You Look: Anchor-Based Evaluation Hides Text Fabrication in Image-Conditioned Video Generation

cs.CV ·2026-10-18 ·Yujia Wang
code/data
Abstract · 18 pages, 2 figures; includes supplementary material (S1-S11). Code and data: https://github.com/geoywang/refanchor-bench

Image-conditioned video models are used to show real places, so whether their output stays faithful to the place they were shown is a practical question. The text benchmarks we found that check fidelity against a reference do so at an anchor - an annotated region, typically the sign the task is about. We show that this choice hides most of the failure. We conditioned four models from three vendors on 48 street-level references of Tokyo storefronts and measured text fidelity at the annotated sign and in the periphery, the other text the reference covers. At the anchor, all four models are close to a no-model control (anchor text fidelity 0.90-0.94 against 0.996) and are not reliably separated from one another. In the periphery, 36-54% of generated text regions have no counterpart in the reference, against 1.6% for the control, although every prompt forbade new readable text. Two blinded raters agree on which regions are fabricated (kappa = 0.83-0.96), the result held in preregistered rounds on two vendors whose videos the measure had never seen (42% and 56%), and VBench-I2V's reference-based consistency scores show no evidence of tracking it in our sample. Most fabricated text is legible-looking text where the reference shows none or cannot be read - a form that string-matching metrics discard, which is why our own first measure missed it. A preregistered Latin-script replication in London missed its bar in a first round and met it in an independent second round. We release code, annotations, per-video scores, ratings and every preregistration. Code and data: https://github.com/geoywang/refanchor-bench

0.47

The Bottleneck Is Retrieval, Not Reading: Diagnosing Evidence-Based Generation with RAG for Long Context Question Answering

cs.CL cs.IR ·2026-10-18 ·Debopriyo Banerjee, Owais Aijaz, Ahmer Tabassum, Zhuohan Xie, Preslav Nakov +1 more
top venuecode/data
Abstract

Retrieval-augmented generation (RAG) is the dominant paradigm for grounding large language models in external corpora, but its similarity-driven, chunked pipeline has structural blind spots: it cannot anticipate which passages a question requires and discards document-level context that resolves cross-passage references. We introduce evidence-based generation (EBG), requiring models to return an answer with line-level citations, and a joint Ans+Evid metric that scores correctness and grounding simultaneously. We reformulate two corpora into evidence-grounded multi-document QA benchmarks- MultiHop-RAG and QASPER-MP, a multi-paper adaptation of QASPER with realistic short paper cues, and pair them with a controlled synthetic two-hop haystack and an evidence-augmented NoLiMa spanning 10K-1M characters. Across ten proprietary frontier and three open-weight LLMs, EBG isolates two RAG limitations that long-context models avoid: similarity retrieval surfaces any relevant chunk far more reliably than all required chunks for multi-fact queries, and chunking destroys document meta-information needed to resolve identity-bearing queries. EBG also exposes grounding failures invisible to answer-only evaluation, including incomplete citations in two-needle settings and a widening gap between Ans and Ans+Evid when context contradicts parametric knowledge. Data and code are available at https://github.com/debobanerjee/LC-EBG.

0.47

Holo4D: Holistic 4D Reconstruction as Geometric Control for Video Diffusion

cs.CV ·2026-10-18 ·Yushi Lan, Zeren Jiang, Koichi Namekata, Xingang Pan, Chuanxia Zheng +1 more
top venuecode/data
Abstract · Accepted to NeurIPS 2026. Project page: https://nirvanalan.github.io/projects/holo4d/

Synthesizing a monocular video from a new camera trajectory requires preserving the observed scene geometry and motion while completing regions that become newly visible. Feed-forward 4D reconstructors recover the structure of the observed scene, whereas video diffusion models provide a generative prior for completing unseen content. We introduce Holo4D, a video reshooting framework that uses reconstructed 3D trajectories as a holistic source-view 4D scaffold. We project these trajectories into the requested views and encode their geometry and source appearance to guide a pretrained video diffusion model. Together with the source video and target cameras, the scaffold grounds generation in reconstructed observations while leaving unsupported regions to the generative prior. Experiments demonstrate the value of trajectory-based conditioning for camera control and novel-view synthesis. The framework accommodates different 4D reconstructors without retraining the video generator, allowing it to benefit from future advances in source reconstruction. More details can be found in our project page.

0.47

PRET is a few-shot system for pan-cancer recognition without example training

cs.CV ·2026-10-18 ·Yi Li, Ziyu Ning, Tianqi Xiang, Qixiang Zhang, Zhihao Lin +15 more
publishedhighly cited
Abstract · Nature Cancer Accepted

Pathological examination stands as the cornerstone in cancer diagnosis, impacting millions worldwide annually. With the shortage of pathologists globally, artificial intelligence (AI) has emerged rapidly to automate the diagnostics process. However, conventional AI models require substantial labeled data for each disease, posing huge challenges in scalability and practicality. Therefore, we introduce PRET (pan-cancer recognition without examples training), a few-shot system to achieve flexible, scalable, and effective cancer recognition across diverse organs, hospitals and tasks without training. Evaluated on 23 international benchmarks comprising 4,484 whole-slide images, our method outperforms existing approaches across 20 tasks, achieving over 97% area under the curve on 15 benchmarks with a maximum improvement of 36.76%. Notably, PRET delivers clinical-grade diagnostic performance in lymph node metastasis detection using only eight slide examples, outperforming 11 pathologists. By offering a flexible and cost-effective solution for pan-cancer recognition, PRET paves the way for accessible and equitable AI-based pathology systems, particularly benefiting minority populations and underserved regions.

0.46

Painting Change Map as guidance for generating painting processes

cs.CV ·2026-10-18 ·Markus Pobitzer, Chang Liu, Niccolò Biondi, Nicu Sebe
positioncode/data
Abstract

Digital art generation has undergone a massive paradigm shift, moving from static image synthesis to dynamic video generation that can mimic the human painting process. While existing models can synthesize visually stunning time-lapses of paintings being created or "unpainted," they largely treat the creative process as a black box. Current conditioning mechanisms are either too rigid, relying on static geometric cues like Canny edges or depth maps, or too holistic, utilizing simple scalar progress variables that offer no localized control. Consequently, users lack the agency to steer how a painting evolves over time. In this work, we argue that true control over the video generation of artwork requires a joint understanding of two fundamental dimensions: where an artist acts, and when those actions take place. To achieve this, we introduce the Painting Change Map, a novel spatio-temporal control mechanism that explicitly tracks normalized structural changes across a 1,000-timestep continuum, a resolution sufficient to capture fine-grained stroke-level changes while remaining computationally tractable. By mapping localized visual differences into an encoded RGB space, our representation enables fine-grained guidance for both the forward painting and reverse unpainting processes. Extensive benchmarking demonstrates the efficacy of our approach. We outperform existing control baselines and achieve competitive or superior results compared to state-of-the-art painting methods. Furthermore, under oracle conditions, our method surpasses all baselines by a significant margin. Code is available at https://github.com/Markus-Pobitzer/painting-change-map.

0.45

Test-Time Scaling of Text-to-Image Flow Matching Models with Noise-Rejuvenated Sequential Monte Carlo

cs.CV ·2026-10-18 ·Wenyi Lian, Zheng Zhao, Thomas B. Schön
publishedcode/data
Abstract · Accepted to TMLR

Test-time scaling of diffusion models with sequential Monte Carlo (SMC) has become a promising strategy for sampling under user-specified rewards, such as human preference scores. However, the standard SMC framework typically relies on stochastic proposals and therefore does not directly transfer to flow matching models, in which inference is governed by deterministic ODEs. In this work, we introduce noise-rejuvenated SMC, an inference-time scaling method tailored for text-to-image flow matching models. Our key idea is to perturb each resampled particle by changing its source noise and constructing a new intermediate state using its current endpoint estimate. This moves the particle to another state that remains representable under the flow-matching interpolation form, enabling particle exploration without converting the sampler to an SDE. We further introduce a look-ahead mechanism to refine resampled particles before weighting. Experiments on Stable Diffusion 3, FLUX.1-dev, and Z-Image demonstrate consistent improvements in reward-guided generation, while ablation studies validate the contributions of noise rejuvenation and look-ahead refinement. Code is available at https://github.com/shermanlian/NR-SMC.

0.45

Calibration-Free Early Stopping for Chain-of-Thought Reasoning via E-processes

cs.CL cs.AIcs.LG ·2026-10-18 ·Dong-Jae Lee, Gyojin Han, Junmo Kim
top venue
Abstract · EMNLP 2026 Findings

Reasoning large language models (LLMs) benefit from test-time scaling, motivating adaptive policies that spend tokens when they help and stop early otherwise. Existing rules halt when a confidence or stability signal crosses a threshold, but such thresholds rarely transfer across models and datasets, and validation-calibrated rules lose their guarantee under shifts in test-time horizon or signal dynamics. We reframe adaptive stopping as sequential evidence accumulation along the reasoning trajectory rather than threshold calibration on held-out data. Our framework runs two e-processes on a shared per-chunk signal: a commit stream tracks the signal's position relative to its slow trend, and an abstain stream tracks its signed velocity, making the two anti-correlated by construction. Under signal-level nulls on the per-chunk increments, Ville's inequality yields an anytime-valid signal-level false-alarm bound at user-specified significance levels, with no held-out calibration. Across various model--dataset combinations, the framework matches or exceeds calibration-based stopping on the token--accuracy Pareto front using zero calibration data, and the same significance level transfers unchanged across cells where calibration becomes infeasible.

0.43

SZED at ArGuard Shared Tasks: Multi-Task Label-Attention and Cross-Modal Fusion for Arabic Hateful Meme Detection

cs.CL cs.AIcs.CVcs.LG ·2026-10-18 ·Selma Khelili, Zohra Abdeli
published
Abstract

We present SZED's submission to ArGuard 2026 Track A: Arabic hateful meme detection. The task has two parts: binary classification (A1) and fine-grained multi-label tagging (A2). For A1, we fuse MARBERTv2 and CLIP ViT-L/14 embeddings, comparing concatenation, self-attention, and ensemble heads. For A2, we use a multi-task label-attention architecture with per-class query vectors and an auxiliary head for cascade calibration. We reach macro-F1 = 0.7597 on A1 and 0.3142 on A2 (averaged over 10 active classes). We report reproducibility across runs, per-class error patterns, and the effect of cascade calibration on the long-tailed A2 label space.

0.43

Multimodal-Conditioned Latent Diffusion Models for Spatiotemporal Brain MRI Forecasting

cs.CV cs.AI ·2026-10-18 ·Antonio Scardace, Silvia Ingala, Sebastiano Battiato, Daniele Ravì
publishedcode/data
Abstract · Accepted to WACV 2027

Spatiotemporal forecasting of 3D anatomical changes is a critical generative vision challenge for modeling Alzheimer's Disease (AD) progression. While brain MRI forecasting from single-modality data has advanced, other key complementary multimodal biomarkers remain underutilized. In this work, we propose the Multimodal-Conditioned Latent Diffusion Model (MC-LDM), a novel framework for patient-specific brain MRI synthesis. MC-LDM leverages multimodal biomarkers---MRI, tau-PET, Clinical Dementia Rating (CDR), and APOE4---during training, but relies only on baseline MRI, APOE4, and patient metadata at inference, reflecting realistic clinical availability. Our contributions are threefold: (1) a Latent Diffusion Model (LDM) fine-tuned via Low-Rank Adaptation (LoRA) for data-efficient learning; (2) a multimodal conditioning that integrates molecular and clinical signals to model heterogeneous disease trajectories; and (3) auxiliary predictors estimating unobserved multimodal biomarkers from MRI features and patient covariates to condition the diffusion model, removing the need for these assessments at inference. Trained on 1,355 visits from 852 subjects in the Alzheimer's Disease Neuroimaging Initiative (ADNI), MC-LDM reduces average regional volumetric MAE by 18% over the second-best baseline on an independent test set of 664 visits. To our knowledge, MC-LDM represents the first diffusion-based framework to leverage multimodal priors for MRI forecasting. This conceptual advancement enables patient-specific trajectory generation that captures aspects of AD heterogeneity not observable from structural anatomy alone, bridging the gap between rich multimodal supervision and the strict constraints of routine clinical deployment. Code is available at: https://github.com/antonioscardace/MC-LDM.

0.43

Epi: Towards a Persistent, Self-Improving Agent for Autonomous Research

cs.CL cs.AI ·2026-10-18 ·Yue Huang, Zhangchen Xu, Xiaonan Luo, Zheyuan Liu, Lichi Li +14 more
benchmarkcommunity pick
Abstract · 23 pages, 7 figures. Technical report

We present Épi, an autonomous research agent that sustains investigations across model sessions and evolves its methods through accumulated experience. Épi combines persistent research history, hierarchical memory, branching and independent review to maintain research continuity, learn from experiments, and explore and revisit alternative approaches. Its experiments run across session boundaries, and a fixed verifier evaluates their outputs. Épi produced 120 record-setting results, comprising 20 numerical bound records across 12 families and 100 construction records across eight families. These include an approximately 84-fold narrower interval for the $n$-queens constant and a more than 3.6-million-fold narrower rigorous interval for the quartic Feigenbaum attractor dimension. Exact dual certificates establish that the Cohn-Elkies linear-programming bound cannot be sharp in dimensions 10 and 11. Beyond numerical improvements, Épi gives a computer-assisted proof of Tuza's conjecture for split graphs with an eight-vertex clique part and at most three triangle-active neighborhood types, with arbitrary multiplicities. We extend this research process to algorithmic coding with Épi-turbo, a lightweight configuration that develops executable solvers for tasks including vehicle routing, scheduling and production planning. On 30 FrontierOR tasks, Épi-turbo achieves a 5.4% higher total benchmark score than Codex at 64% of its model cost, using the same underlying model and reasoning effort. Épi is designed to support human understanding by connecting computational results to explicit methods and mathematical arguments that researchers can examine and build upon.

0.43

Legal Guidance in the Wild: How Users Seek Legal Help in Real-World LLM Conversations

cs.CL cs.CY ·2026-10-18 ·Amirreza Mohseni, Gerasimos Spanakis
publishedcode/data
Abstract · 12 pages, 6 tables. Accepted for publication at JURIX 2026, Toulouse, France. This is the preprint version. Code and data available at: http…

People ask conversational AI models about their legal rights, obligations, procedures, and personal situations that may have legal consequences. To study this behavior, we first need to distinguish requests for legal guidance from other legally themed discussions and unrelated conversations. We contribute a multi-axis taxonomy of legal queries capturing whether legal guidance is sought, how a request is formulated, and which legal domains it concerns; apply it to 2,122 naturally occurring WildChat-1M conversations; and create a human-reviewed gold evaluation set of 200 conversations. All annotations, model checkpoints, and evaluation code are publicly available. In this enriched sample, 71.1% of conversations containing legal content seek personalized or practical guidance. Rights-and-options advice is the most common primary request type among legal-content conversations, and 25.3% of guidance-positive conversations span multiple legal topics. The gold evaluation set supports a two-stage comparison of ModernBERT-base and ModernBERT-large against a TF-IDF cascade. Our framework supports large-scale studies of legal guidance in conversational AI, and the public release of the annotated data facilitates reproducible research and future model development.

0.43

HistGramBench: Measuring Structural Recovery When Multimodal Models Redraw Historical Mathematical Diagrams as TikZ

cs.CV cs.CLcs.DL ·2026-10-18 ·Minh Quang Tran, Carter He, Cameron Bloch, Zitai Huang, Thien Truong Vo +2 more
benchmarkcode/data
Abstract · 34 pages, 8 figures, 12 tables. Code: https://github.com/VietAlpha/histgrambench ; data: https://huggingface.co/datasets/VietAlphaLabs/HistG…

Digital libraries preserve mathematical books printed from the sixteenth to the twentieth century, but the lettered diagrams in them, the figures on which Euclidean proofs depend, remain scanned images whose content cannot be searched. Redrawing such a diagram as a short drawing program would record which points lie on which lines and circles. But a printed plate offers no source code against which the redrawing could be checked. We introduce HistGramBench, a benchmark of 2,507 diagrams from 95 works printed between 1551 and 1933 in eight languages. For each diagram, two AI models from different providers agree on the facts a reader takes from the figure, such as which lines are parallel or where two lines meet, and every redrawing is checked against those facts on the model's own drawing. We evaluate six affordable AI models. The best recovers about three fifths of these facts, and most of what the models miss comes from programs that never produce a drawing. Parallels and tangencies are the hardest relations to reproduce. HistGramBench gives historians and librarians a way to measure when such tools can make old mathematical diagrams searchable.

0.42

PRIMA: Pose Refined Inductive Motion Adapter for Zero Shot Social Bodily Behavior Recognition

cs.CV cs.AI ·2026-10-18 ·Mingcong Ding, Babette Bühler, Tim Fütterer, Ulrich Trautwein, Enkelejda Kasneci
top venue
Abstract · 10 pages, 5 figures. Accepted at ACM Multimedia 2026 (ACM MM 2026)

Recognizing fine-grained bodily behaviors, such as self-grooming, postural shifts, and backchanneling, is a useful intermediate step for computational social signal analysis. But the behaviors that matter vary across social contexts, and retraining a model for each new taxonomy is expensive. Vision-language models (VLMs) can adapt to new categories through text prompts, yet they lack sensitivity to body structure and often confuse subtle postural distinctions. We propose PRIMA (Pose Refined Inductive Motion Adapter), a lightweight module that injects a body-structure prior into frozen VLMs using human pose as a training-time supervisory signal, requiring no pose input at inference. PRIMA separates a model-specific projection head from a shared core that encodes transferable body-configuration knowledge; adapting to a new VLM requires retraining only the head. We train PRIMA on a micro-action dataset and evaluate entirely through zero-shot prompting. The adapter improves recognition on its training domain, transfers to additional VLM backbones at different scales with the core frozen, and generalizes to two unseen target datasets from social psychology and daily activity recognition. In a supporting downstream study, the improved behavior predictions also yield higher correlations with human ratings of rapport and dominance in a collaboration context. Within the scope of the tested models and indoor datasets, the results indicate that pose-guided training can produce body-structure priors that generalize across unseen behavior vocabularies and can be partially reused across VLM backbones while preserving RGB-only inference.

0.41

One Head Away from Language Confusion: Single Attention Heads Across Model Families

cs.CL cs.AI ·2026-10-18 ·Jiwon Jeong, Anjelo Jann Laroza, Chaewoon Yoon, Arjun Pillai, Pranav Kokati +2 more
benchmarkcode/data
Abstract

Large language models sometimes reply in an unexpected language, a failure known as language confusion. We ablate individual attention heads in 12 model checkpoints spanning seven model series, using FLORES-200 continuations in five languages and Language Confusion Benchmark (LCB) chat replies. In ten models, zeroing one head causes substantially more departures from the expected language than zeroing any other head in its layer; its LCB pass-rate drop persists, though reduced, under mean ablation. Seven heads move 8.0 to 51.5 percent of all FLORES continuations out of the prompt's language, versus at most 1.4 percent for other same-layer heads; the other three mainly affect explicitly requested languages. These seven heads matter less in base models given plain text. In two models, replacing the values a head reads at the requested-language name with another language's mean makes replies fail and shifts some into that language, unlike a random change of equal size. Replacing a head's output with another language's mean redirects 84 to 99 percent of scored non-English monolingual LCB replies into that language in five models. Yet the requested language's mean improves crosslingual replies beyond every control head only in the two Gemma models, whose baselines are low. Code and results are available in the public repository at https://github.com/christian-hoang-04/anonymous-release.

0.41

Reflection-Guided Dynamics Rectification for Text-to-Image Generation

cs.CV cs.AI ·2026-10-18 ·Shuo Xu, Hanrui Ying, Debin Liu, Jianjun Li, Zhiyuan Ma
code/data
Abstract · This paper is 20 pages long, including the appendices, and contains 10 figures. Interactive demos, code, and the dataset are available on th…

Fine-grained rectification of existing generations under a limited additional budget remains a key problem in test-time optimization (TTO) for image generation. As for prior approaches, they could identify residual errors, but could not directly ensure that the model acts on the feedback during reasoning. Such errors stem from multiple sources, such as the initial noise or text prompt, and ultimately induce an execution gap of failing to handle potential deviations along the trajectory generating. However, classic models generally cannot remedy this gap. To narrow it, we propose Reflection-Guided Dynamics Rectification (RGDR), which turns reflections on completed outputs into corrective interventions on the process of sampling dynamics at the trajectory level. Specifically, RGDR could impact every sampling step by adapting to instantaneous intervention. Moreover, RGDR generates flexible candidates while adjudicating between them and the reference. As experiments show, rectification from RGDR starts from the existing base trajectory and reflection, and RGDR is substantially more effective than former trajectory exploration. Under matched budgets for additional image generation, RGDR's gain in the cross-model average final checklist score over the reference is $1.51\times$ that of TIR prompt refinement and $1.81\times$ that of Dynamic CFG.

0.40

From Unverifiable Domain Text to Cross-Lingual RLVR via Code-Switched Task Synthesis

cs.CL ·2026-10-18 ·Cheonbok Park, Jihwan Kim, Shinyoung Joo, Joosung Lee, Hwiyeol Jo +2 more
top venue
Abstract · Accepted in EMNLP2026

Reinforcement Learning with Verifiable Rewards (RLVR) can improve reasoning in large language models, but specialized non-English domains expose a data mismatch. Raw English domain text is often available, while target-language RLVR tasks with verifiable rewards are scarce. We introduce code-switched task synthesis for cross-lingual RLVR, a data-centric approach that derives target-language and code-switched RLVR task variants from otherwise unverifiable English domain text. Using fill-in-the-middle multiple-choice QA as a seed task format, we synthesize automatically verifiable tasks from domain passages and reshape them into cross-lingual variants. These variants preserve the same evidence-answer relation but change how English and target-language evidence are presented during RLVR. On Korean medical benchmarks and cybersecurity benchmarks across four target languages, code-switched training outperforms matched monolingual and bilingual-mixture baselines on average target-language performance while maintaining comparable English performance. Additional analyses provide evidence of better target-language input understanding and stronger cross-lingual alignment in both model outputs and internal representations. These results suggest that code-switched training helps target-language inputs access domain reasoning learned more strongly in English.

0.39

Structure-Aligned History Learning with Subspace-Structured Semantic IDs for Generative Recommendation

cs.IR ·2026-10-18 ·Kaiwen Wei, Kejun He, Jie Zhang, Xiaomian Kang, Yuming Yang +10 more
code/data
Abstract · 34 pages, 10 figures, including references and appendices

Generative recommendation constructs semantic IDs for catalog items and predicts the next ID from user interaction histories. However, existing methods face limitations in both stages. First, conventional quantizers may produce imbalanced code assignments and lack coherent intermediate semantic units between individual tokens and complete items, limiting history reconstruction. Second, recommendation models remain dominated by next-item prediction, treating historical interactions mainly as conditioning context and providing limited supervision for recovering persistent preferences across the historical path, particularly when relevant preference signals are distributed over distant interactions. To address these limitations, Structure-Aligned History Learning (SAHL) reconnects the two stages by using the structure created during quantization directly for history reconstruction. Its Semantic Residual Balanced Quantization (SRBQ) forms a balanced token-subspace-item hierarchy, and SAHL reconstructs history at these same granularities. Predictive entropy prioritizes uncertain transitions, while curriculum optimization progressively shifts training to inference-aligned next-item prediction. On four Amazon Review categories, SAHL improves NDCG@10 by 18.6%-30.0% over RPG and 2.7%-6.0% over strong method MHL. Controlled analyses verify balanced code use, the benefit of following quantizer-defined groups during reconstruction, and dependence on user history. The relevant code is available at https://github.com/CQU-MM-Intelligent-Lab/SAHL.

0.38

GeoV2V: Geometry-Grounded Video Diffusion Model for Driving Scene Generation

cs.CV ·2026-10-18 ·Yuchen Xi, Yuhan Guo, Chenwei Hou, Zixu Liu, Jin Fang +2 more
top venuecode/data
Abstract · Published at ECCV 2026. Author accepted manuscript. Project page: https://xiyuche.github.io/GeoV2V/

Novel view synthesis under trajectory changes is essential for autonomous driving simulation, yet existing methods struggle to generate consistent videos when extrapolating to unseen viewpoints such as lane shifts. Reconstruction-based approaches maintain geometric consistency but degrade in visual quality under large viewpoint changes, while reconstruct-then-restore methods often produce visually plausible frames yet suffer from geometric and temporal inconsistencies. We present GeoV2V, a geometry-grounded video-to-video diffusion framework that synthesizes temporally coherent driving videos along novel trajectories by integrating LiDAR and depth-derived geometric priors with full-video conditioning in the Wan 2.1 backbone. A key challenge in learning cross-lane transformations is the lack of synchronized multi-trajectory supervision in real-world datasets. To address this, we also introduce Para4D, a synthetic dataset capturing synchronized multi-lane driving videos within dynamic scenes, which provides explicit supervision for lane-shift generation and serves as a strong training prior for geometry-consistent view synthesis. Experiments on Waymo, nuScenes, and Para4D demonstrate state-of-the-art performance in visual quality, geometric consistency, and robustness to complex lighting dynamics under large trajectory shifts. Project page: https://xiyuche.github.io/GeoV2V/

0.38

Flattening Fantasies: Analyzing Stereotypes in Pornographic Discourse

cs.CL ·2026-10-18 ·Sarmad Chandio, Osama Khalid, Sharaf Zia, Ethan Kutlu
top venueunusual
Abstract

This study presents a large-scale quantitative analysis of 5.4 million video titles from xHamster, the world's fourth most visited pornography platform, to examine linguistic patterns used to describe different dimensions of identities (e.g., race and gender). Specifically, we investigate how nationalities, ethnic groups, and gender identities are portrayed in pornographic discourse, focusing on stereotypes and deviations from implicit norms of pornography. The results show the dominance of American pornography over other groups, which acts as the de facto norm for the adult industry. We identify persistent ethnic stereotypes reflecting colonial power dynamics, religious fetishization, and the sexualization of cultural taboos. The analysis reveals gendered representational differences across cultures, with a prevalent Madonna-Whore dichotomy in the portrayal of women and stereotypical depictions of men in physically dominant roles.

0.37

From Seeing to Reporting: A Theory of Reliable Visual Set Generation in Vision-Language Models

cs.CV cs.CLcs.LG ·2026-10-18 ·Junzhe Shi, Yuan Gan, Shida Jiang
positioncode/data
Abstract · 56 pages, 6 figures. Code, prompts, configuration and all 67,364 stored model responses: https://github.com/shijunzhe/vlm-set-reporting

Vision-language models (VLMs) are asked to report what they see: to count objects, box them, parse a screen or pick a click target (GUI grounding). We argue that reporting is part of the visual task, not merely its formatting. In a dense scene with 173 shapes, GPT-6.1 lists 173 records that match every shape, yet the counts it states in the same response sum to 188. From Seeing to Reporting (FSR) is a theory of visual reporting and a set of inference-time procedures that improve the reports of frozen VLMs. It models reporting as a chain from responses to decoded records, constructed sets and task answers, linked by task-compatible matching. The chain gives exact relations for deleting records, averaging their coordinates and switching readouts, and it splits the error of one answer chosen among several calls into missing candidates, lost representatives and wrong choices. These relations guide procedures for counting, detection, document layout, screen parsing and GUI grounding that use the task specification and a shipped configuration, without requiring labels from the target task. On 50 dense synthetic scenes, a one-call reporting-and-decoding protocol cuts per-type count error by 85.2% for GPT-6.1 and 98.8% for Claude Sonnet 5.5. On 600 ScreenSpot-Pro items, a GUI-grounding tier that combines three models lowers the miss rate from 13.67% for GPT-6.1's two-stage zoom to 9.33%, against 12.67% for three such zooms combined by a cluster vote at about the same cost. A plain vote over the same three models misses 12.33%: the other models supply the candidates, and selecting among them turns those candidates into fewer misses. We release the protocols, the stored responses and the analysis code.

0.36

Deep Auxiliary Learning via Bilevel Feature Optimization for Label-Efficient Medical Image Segmentation

cs.CV ·2026-10-18 ·Fuping Wu, Xi Jia, Le Zhang, Jiahua Li, Rima Mustafa +2 more
code/data
Abstract

Auxiliary learning (AL), which can improve the performance of a primary task by leveraging related auxiliary tasks, is commonly regarded as a special case of multi-task learning. Existing AL approaches typically require manually annotated auxiliary labels, increasing annotation costs, particularly in medical imaging where expert labeling is expensive. In this work, we investigate whether segmentation performance can be improved using auxiliary tasks whose labels are automatically derived from existing segmentation annotations, requiring no additional manual labeling. To address this, we propose a novel auxiliary learning framework, formulated as a bilevel optimization problem. Our method enables effective knowledge transfer between tasks via a feature communication module, enhancing the primary segmentation task. We evaluate the method on three medical image segmentation benchmarks under limited-label settings. The results demonstrate consistent improvements over state-of-the-art methods, achieving superior label efficiency. The code is publicly available at https://github.com/FupingWu90/ALBO.

0.36

Over-Defense in Coding Agents Begins at Silent Forks

cs.CL cs.LGcs.SE ·2026-10-18 ·Hanjun Luo, Xuecheng Liu, Xiucheng Zhang, Zhuoning Xu, Yingbin Jin +3 more
code/data
Abstract · 27 pages, 10 figures, 31 tables. Code: https://github.com/Astarojth/silent-forks

Language models behind coding agents are post-trained with outcome rewards, and recent methods concentrate credit on the turns that decide the outcome. We find that defensive actions beyond what a task requires escape both. They begin at silent forks, the turns where every next action leads to the same outcome. Such actions leave the success rate unchanged, and frontier models take the disproportionate branch there in up to 91% of decisive actions under implicit budgets. The most direct fix, a checker in the trajectory reward, reaches these forks only through fully proportionate trajectories, leaving 74% of GRPO groups with identical rewards and behavior unchanged. This motivates a rethink of where learning signals belong, namely where behavior diverges, which a behavioral checker can locate and the outcome cannot. Our method, ForkDPO, contrasts a proportionate and a disproportionate action at the same prefix, at silent forks on both sides of the required behavior. On a test set of unseen repositories, it lowers a 4B model's disproportionate behavior from 70.3% to 58.9%, against 67.7% for DPO at randomly chosen turns, establishing behavioral divergence as a learning signal for how agents act beyond whether they finish. The code is publicly available at https://github.com/Astarojth/silent-forks.

0.36

Preserving the Manifold: Frozen Expert Alignment for Label-Efficient SAR Scene Understanding

cs.CV ·2026-10-18 ·Eli A. Underwood, I. Taylor McKechnie, Leslie M. Collins
published
Abstract · 10 pages. A shorter version of this work was presented at the ECCV 2026 Workshop TerraBytes II: Towards Global Datasets and Models for Earth…

Cross-modal alignment of synthetic aperture radar (SAR) imagery to natural language is constrained by two distinct failures. The first is modality mismatch: text encoders pre-trained on optical imagery, such as CLIP, cannot be efficiently fine-tuned on SAR, because active microwave backscatter is physically incompatible with passive optical reflectance and the available SAR caption datasets are orders of magnitude too small to bridge this gap without substantially degrading the learned semantic manifold. The second is physical forgetting: SAR encoders pre-trained on large-scale self-supervised objectives encode radar physics that are progressively erased when the encoder is fine-tuned for text alignment on limited caption data. A single design principle addresses both failures: keep both encoders frozen and learn only a constrained, structure-preserving bridge. Our Frozen Expert framework aligns a frozen CROMA SAR encoder with a frozen CLIP text encoder via a lightweight constrained residual bridge. We formally define physical forgetting via Procrustes disparity and show our bridge preserves SAR geometry significantly better than fine-tuning (disparity: 0.210 vs. 0.351). We validate generality across multiple SAR backbones and text encoders. On zero-shot SAR scene classification, our approach outperforms fine-tuned baselines using up to 8x less training data: DFC2020 (+10.3% R@1, +15.5% mAP), EuroSAT-SAR (+15.7% R@1, +12.2% mAP), BigEarthNet-MM (+14.7% R@1). These results establish frozen manifold alignment as a principled, data-efficient alternative to fine-tuning in high-disparity cross-modal settings.

0.35

DiffGround: Semantic Difference-Guided Zero-Shot Video Question Grounding

cs.CV ·2026-10-18 ·Cheng-Han Lee, Amir Salimi, Hsiang-Wei Huang, Che-Chun Su, Min Sun +1 more
publishedcode/data
Abstract · This paper is accepted by ACCV 2026. We plan to publicly release the source code at https://github.com/steven413d/DiffGround

Video Question Grounding (VideoQG) requires both accurate question answering and precise temporal localization. Existing supervised methods rely on costly temporal annotations, while existing zero-shot approaches are primarily designed for Video Question Answering (VQA) and are not optimized for temporal grounding. We propose DiffGround, an action-driven zero-shot framework built upon frozen Multimodal Large Language Models (MLLMs). Our framework first predicts an answer-relevant action cue that transforms the input question into an explicit semantic target for grounding. Action Cue converts a reasoningoriented question into an action-centric retrieval query. Guided by the action cue, a Vision-Language Model (VLM) efficiently selects informative keyframes for reasoning. Finally, we introduce Semantic Diff Refinement, a training-free temporal refinement method that exploits semantic differences between neighboring SigLIP2 embeddings to recover complete action segments from sparse predictions. Extensive experiments on NExT-GQA and STAR show that DiffGround consistently outperforms existing zero-shot and supervised methods, establishing a new state of the art for Video Question Grounding.

0.35

From Zero to OCR: A Data-Centric Framework for Unsupported Low-Resource Scripts

cs.CV cs.CL ·2026-10-18 ·Faizan Iqbal, Haq Nawaz Malik, Hafsah Pervaiz
code/data
Abstract · 9 pages, 9 figures, 4 tables. Model: https://huggingface.co/Faizaniqbal/KoshurOCR-CRNN-Baseline

Optical character recognition (OCR) remains inaccessible to many low-resource scripts because annotated document images are unavailable and synthetic training data can fail to preserve script-specific visual structure. We present a data-centric framework that builds end-to-end OCR for such scripts from only an unannotated text corpus and digital fonts, without manually annotating a single training image. A synthetic data engine renders word, line and page images with correct complex-script shaping and automatically derived labels; a YOLOv8 word detector and a convolutional recurrent recognizer trained with connectionist temporal classification learn from these images; and a stage-gate evaluation protocol isolates recognition error before detection is introduced. We demonstrate the framework on Kashmiri, a Perso-Arabic Nastaliq script with approximately 6.8 million speakers that none of six widely used OCR engines supports. Three line-level recognizers, including one transferred from Arabic, stalled between 41% and 79% character error rate (CER), motivating a word-level formulation; subsequent improvements came primarily from rendering fidelity, diacritic-preserving crop margins and tuned non-maximum suppression. The resulting system achieves 0.39% CER on held-out words, 10.0% on synthetic pages and 13.5% on browser-rendered digital pages, and establishes reference results of 18.45% on typeset book pages and 49.8% on degraded historical documents. We publicly release the trained models, the inference code and a synthetic dataset of 285,864 Kashmiri text-line images, and we describe a step-by-step procedure for building OCR for other unsupported scripts.

0.35

Whistle and Wrong Lyrics: Synthetic Queries Are Not Enough for Real-World Query-by-Melody

cs.SD cs.IRcs.LG ·2026-10-18 ·Luigi Rivolta
code/data
Abstract · 15 pages, 2 figures, 6 tables. Code: https://github.com/luiginotmario/hum2song

People look up a song by humming it, but also by whistling it or singing it with the wrong words. Published query-by-humming systems and benchmarks focus on hums and sung queries, and no public dataset pairs whistles or wrong-lyrics singing with song recordings. We study both query types with a melody-only retrieval system: a key-normalized pitch track embedded by a 5.2M-parameter contour encoder, with a spectral-peak tracker for whistles. First, a physics-grounded generator renders synthetic whistles and nonsense-syllable singing from the vocal melodies of 11,083 songs. Training on them raises top-1 retrieval of held-out synthetic queries against an 11,050-song gallery from 0.180 to 0.251 for whistles and from 0.193 to 0.264 for wrong-lyrics singing, but the gain does not transfer: on the MLEnd test performers the best synthetic-only model gets 210 of 339 real whistles right at top-1, against 214 for the baseline. Second, a closed-set fine-tune on real MLEnd whistles with performer-disjoint splits raises whistle-to-hum top-1 from 0.631 (95% CI 0.581-0.681) to 0.673 (0.622-0.723), on known songs only. Third, two further generations of synthetic queries also fail to transfer: melody-conditioned whistles and la-la singing from a commercial neural music model leave real-whistle top-1 at 0.631, and an open-source pipeline (YingMusic-Singer-Plus singing with partly wrong, mumbled or nonsense lyrics, a DDSP whistle synthesizer, and a melody-error model fitted on real performers) gives at best 0.619 (paired difference -0.012, 95% CI -0.038 to +0.015). Across physics-based, commercial neural and open-source neural synthetic data, nothing transferred to real whistles, which points to collecting real queries as the missing step. Code is public; models and data will be released.

0.35

DeGMatch: Target-Aware Reasoning Data Selection via Deconfounded Gradient Matching

cs.CL ·2026-10-18 ·Changho Lee, Janghoon Han, Rajdeep Mukherjee, Yeda Song, Lening Nick Cui +2 more
code/data
Abstract

Reasoning post-training on long trajectories has substantially improved the reasoning capabilities of large language models, but scaling such training to large datasets is increasingly expensive. We study target-aware reasoning data selection, where a small reference set specifies a target capability and guides the selection of training trajectories for specialization. We find that existing target-aware methods can exhibit limited target discrimination on long reasoning trajectories, while dominant population-level structure in candidate gradients can obscure finer target-dependent variation during gradient matching. To address these limitations, we introduce Deconfounded Gradient Matching (DeGMatch), which first constructs a target-relevance gate and then reranks candidates using model-specific gradients. DeGMatch centers gradients using the candidate-population mean and removes leading principal directions before target matching, reducing the influence of dominant population-level structure and improving target discrimination. Across mathematics, science, and code, DeGMatch outperforms strong heuristic, target-aware, and reasoning-specific data selection baselines, while remaining effective across selection budgets, model families, and scales. Our code is publicly available at https://github.com/CHLee0801/degmatch.

0.34

Acharya 2.0: An AI Tutor for Self-Learning Subjects in Hindi

cs.CL ·2026-10-18 ·Shantipriya Parida, Debasish Dhal, Sonal Khosla, Sambit Sekhar, Muhammed Abdur Rahmaan Kamaldeen
hot direction
Abstract

Text processing is crucial for low-resource languages as it enables the development of tools and technologies that bridge the digital divide and provide equitable access to information and services. By empowering low-resource languages with robust text processing, we can preserve linguistic diversity, enhance education, improve communi- cation, and foster inclusivity in a rapidly digitizing world. Due to the lack of sufficient annotated datasets and lin- guistic resources, it is challenging to develop accurate and reliable language technologies for these low-resource languages. Large Language Models (LLMs) have attracted sig- nificant attention with their performance and the value they produce. However, with the ever-evolving language, adapting LLMs for a specific task is time-consuming and resource-intensive. Thus, Retrieval-Augmented Genera- tion (RAG) has overcome this limitation by augmenting LLMs with external knowledge, thereby improving their efficiency and extensive task-specific adaptation. Most applications of the RAG technique have been limited to English and has seen limited application in In- dic languages. This paper introduces the RAG-powered AI-Tutor application, Acharya 2.0, designed to produce academic assistance to Hindi-speaking students. The client-server-based web application integrates many fea- tures to support students in accessing and interacting with educational content.

0.34

Think Visually Across Space and Time: Grounded Reasoning for Video Segmentation

cs.CV ·2026-10-18 ·Junwei Wen, Deshui Miao, Guangming Lu, Ming-Hsuan Yang, Xin Li +1 more
Abstract · 21 pages

Video Reasoning Segmentation (VRS) requires a model to identify an implicitly described target and predict its pixel-level mask trajectory, while resolving the spatial relations and temporal events that determine the answer. Text-only Chain-of-Thought (CoT) can explain a target choice without making the underlying visual references explicit, leaving a \emph{video reference gap} between linguistic inference and the evidence distributed across frames. We introduce \textbf{Temporal Chain of Video Primitives (TCoVP)}, a spatiotemporally grounded form of Video CoT that organizes query-relevant observations into a chronological reasoning chain. Each primitive associates an entity reference with temporal support and timestamped spatial evidence, and the chain interleaves these references with language to guide reasoning about which entity is relevant, when the queried event holds, and where its evidence appears. The selected target primitives then provide box prompts and temporal support for video mask prediction. We construct TCoVP-SFT-15k and TCoVP-RL-13k to learn this representation through supervised fine-tuning and chain-aligned reinforcement learning. Extensive experiments across five VRS benchmarks and three model backbones demonstrate the strong performance of TCoVP, highlighting the effectiveness of organizing spatial and temporal evidence within the reasoning process. All models and code will be publicly released. Project page: \url{https://victor-wjw.github.io/TCoVP/}

0.34

How Far Can the Gemma 4 Speech Encoder Go Without Changing Its Weights? Saudi and Gulf Arabic Speech Recognition with a Frozen and an Adapter-Augmented Encoder

cs.CL cs.SDeess.AS ·2026-10-18 ·Amr Abdelkhalik
code/data
Abstract · 25 pages, 15 tables, 5 figures. v2: corrected description of the encoder layer mix and other wording fixes; all results unchanged. Code and …

We study the 305M-parameter audio encoder shared by Gemma 4 and EmbeddingGemma 2 (bit-identical to that of Gemma 4 E2B except for the final audio-embedding projection) for Saudi and Gulf Arabic speech recognition, without changing any of its pretrained weights; the adapters of our adapted system are additive and kept separate. A frozen system trains only a 22.6M-parameter CTC reader over a learned mix of four intermediate encoder layers and the encoder output; an adapted system adds 3.1M low-rank adapter parameters, a 41.5M-parameter hybrid CTC/attention reader and joint decoding with a sub-word language model. Against Cohere Transcribe Arabic (2.07B parameters), re-run on identical audio, the frozen system trails in word error rate (WER) but is not significantly different in character error rate (CER) on five of six human-labelled conditions, while running 3.4 times faster on CPU with greedy decoding. The adapted system is not significantly different in WER on a public Saudi customer-service call (19.20% against 19.27%) and on clean Emirati speech, is significantly better on telephone-coded Emirati speech, has a significantly lower CER on all six conditions of these sets, and beats a 0.6B recogniser fine-tuned on SADA in WER on all six; it still trails Cohere in WER on long-form Saudi speech. On a blind, human-corrected public call-centre sample it ranks 4 of 15 systems, behind three commercial APIs (significantly only one) and ahead of the other self-hostable systems by point estimate. An optimised joint decoder runs it at 139 times real time on one GPU. We also find, on our data, that references made by correcting machine drafts favour the drafting system. Code and public-set results are released.

0.34

DF-PALM: Dynamic Fusion Prompt Learning for Audio--Language Models

cs.SD cs.LGeess.AS ·2026-10-18 ·Yulin Sun, Shenghua Li, Yong Dou, Xinwang Liu, Cheng Yang +1 more
benchmarkcode/data
Abstract · 13 pages, 5 figures, 14 tables

Few-shot audio classification requires audio--language models to distinguish variable acoustic events from limited labeled data. PALM's main formulation adapts encoded class representations without changing pretrained encoders, but its class representations remain input-independent and classification uses audio features without class-conditioned refinement. We introduce DF-PALM, a dynamic feature-space prompting framework that combines audio-conditioned class representations with asymmetric cross-attention for audio refinement. Class representations are conditioned after text encoding, held fixed across fusion layers for each input, and reused for final scoring. Both pretrained encoders remain frozen, avoiding audio-conditioned text-encoder passes. With PENGI as the frozen backbone, the main 16-shot evaluation covered 11 benchmarks and six baselines. DF-PALM achieved 80.69% mean accuracy and 80.32% macro-averaged F1, exceeding PALM by 4.11 and 4.30 percentage points, respectively. Paired benchmark-level tests supported both gains after Holm correction. Across four shot settings, DF-PALM achieved the highest mean accuracy and macro-averaged F1 among the four evaluated prompt-learning methods. The gains extended to LAION-CLAP and OpenFLAM, with significant improvements over PALM in both metrics. At 50 candidate classes, operator-counted inference floating-point operations increased by 0.39% relative to PALM. These results support instance-conditioned feature-space adaptation as an effective approach to few-shot audio classification. Our code is publicly available at GitHub https://github.com/sunyulin0421/DF-PALM.

0.34

PanoDent-KD: From Generalist Knowledge Distillation to Cross-Dataset Specialist Detection and Embedded Deployment in Panoramic Radiographs

cs.CV cs.PF ·2026-10-18 ·Muhammad Waqas Ashraf, Imad Ali Shah, Ali Hassan
published
Abstract · Accepted at WACV 2026

Reliable panoramic-radiograph detection requires consistency across datasets with different class definitions and deployment software. We present PanoDent-KD, an auditable 1280-pixel YOLO26m pipeline linking a 12-class source generalist, feature-level knowledge distillation (KD), exact classifier-row transfer, teacher-free five-class specialization, and embedded deployment. A common training and evaluation pipeline compares three source-training arms across three seeds: a supervised baseline (B1), single-teacher KD (STKD), and multi-teacher KD (MTKD). Five shared classifier rows initialize nine OralXrays-9 specialists, while the backbone, FPN-PAN neck, and dual-detection-head topology are retained unchanged. On the 2,686-image test subset, all three arms achieve macro box F1 near 0.82. B1 achieves the highest three-seed arm-level mean macro box F1 (0.8239$\pm$0.0027) and $mAP_{50:95}$ (0.6130$\pm$0.0067). STKD seed 1, selected after development but before final testing, was used for deployment. Its Jetson Orin Nano Super FP16 TensorRT engine achieves 77.03-ms median and 80.86-ms P95 prediction-call latency, throughput of 12.90 images/s, and gross total-module energy of 1.39 J/image, with high numerical agreement between within-device PyTorch and TensorRT FP16 inference. These findings demonstrate the feasibility of an auditable, device-specific deployment workflow, but do not establish clinical validity or improved target-domain accuracy from KD.

0.34

A 3D-Driven Framework for Moving-Camera Perspective-to-360° Video Generation

cs.CV ·2026-10-18 ·Michael Wakeham, Mrinmoy Kumar Das, Wenqi Wang, Ayush Tewari, Boqing Gong +1 more
publishedcode/data
Abstract · Accepted to WACV 2027. Project page: https://pano3d-video.github.io/

Turning a moving-camera perspective video into a 360° video means generating content outside the input frame under several constraints: observations revealed over time must be preserved, unseen regions plausibly completed, and the resulting scene kept consistent as the camera viewpoint changes. Existing methods assign these different requirements to a single conditional video model, requiring them to be jointly learned from panoramic video data. We instead present a modular, 3D-driven framework that distributes these three responsibilities across specialized components, all connected through a shared scene representation. Input observations are reconstructed explicitly in 3D, while a pretrained panoramic image model completes unobserved content from a selected panoramic render. This generated content is lifted into the shared 3D scene, where it remains spatially anchored across camera viewpoints. Without requiring any additional panoramic-video training, our framework achieves state-of-the-art input-view fidelity, perceptual quality, and geometric consistency on a static-scene subset of the Argus evaluation set. These results support our decomposed, 3D-driven formulation as a promising alternative to direct video generation in this setting.

0.33

Hidden Cartographer: Hidden State Guided Data Selection for Dual Budgeted Instruction Tuning

cs.CL ·2026-10-18 ·Zixuan Qin, Qi Qin, Erbo Li, Ting Wei, Yifan Sun +1 more
Abstract

Given the high cost of expert annotation and limited fine-tuning resources, effective data selection from large unlabeled pools under strict budgets has become a crucial step in adapting LLMs to specialized domains. Existing label-dependent methods risk distorted selection when using pseudo-labels, while traditional active learning methods do not explicitly characterize instruction-tuning dynamics. To address these limitations, we propose Hidden Cartographer (HC), a two-stage framework that uses hidden states to guide annotation and exact training dynamics to select the final training subset. The first stage is Progressive Candidate Expansion, which utilizes the proxy to iteratively select top unlabeled candidates for expert annotation, merging them to recompute training dynamics and update the proxy until the annotation budget is exhausted; the second stage is Exact Cartography Pruning, which ranks these annotated candidates by their exact dynamics scores to isolate a compact subset under the final training budget. Experiments across the medical, legal, and financial domains demonstrate that HC achieves the highest average accuracy among all budgeted baselines, requiring annotation for only 20--30% of the original pool and fine-tuning on just 10--15%.

0.33

Quality-Controlled Reply and Summary Data for Post-Training Small Language Models in AI Call-Screening Agents

cs.CL ·2026-10-18 ·Jiaqi Gan, Tsang Ng, Ankit Raj, Tommy Duong, Xingyu Shen +3 more
Abstract · 17 pages, 3 figures, 12 tables

Small language models have improved to the point where a phone can plausibly screen its own calls, answering on an absent owner's behalf without sending call content off the device, a privacy-respecting alternative to cloud call handling. Realizing that promise takes a model that can do two things well: reply within the owner's authority during the call, and afterward tell the owner what happened and what needs attention. Released small models do neither reliably - they raise unwarranted alarms about ordinary callers and write summaries that drop material conditions or report a caller's claims as fact - and real calls supply no reference replies or summaries to learn from. We construct supervision for both tasks from synthetic calls under an information-boundary constraint: a caller writer works from a private scenario brief, a reply writer sees only the conversation up to the current turn, and a summary writer sees the completed transcript. Separate model-based reviews of the full dialogue and the summary decide which calls enter training, yielding a corpus of 3,395 calls with 15,102 targets (11,707 reply, 3,395 summary). Training a single shared LoRA adapter on both tasks, we evaluate six models (0.6-4B parameters) before and after adaptation on 192 held-out calls. Reply qualification rises from 21-54% to 94-97% and summary qualification from 0-6% to 43-72%; all twelve paired gains (six models x two tasks) have bootstrap intervals above zero, and the summary gains hold in every source family. Because one model both screened the training data and scored the outputs, these rates measure agreement with that evaluator's standard rather than human acceptance, which a planned human study will test. After adaptation, coherence and naturalness failures nearly vanish, leaving grounding and usefulness - faithfully tracking who said what, under which conditions - as the remaining challenge.

0.33

SlimFormer++: Token Mixing Where It Matters for Efficient 3D Medical Image Segmentation

cs.CV ·2026-10-18 ·Yang Hong, Lei Zhang, Zhihong Liu, Jianqing Mo, Xujiong Ye
code/data
Abstract

To determine where computation is needed in volumetric segmentation networks, we analyze block-level redundancy and stage-wise token-mixer contribution. Centered Kernel Alignment reveals high input-output representational similarity within blocks of representative convolutional, hybrid, and Transformer-style segmentation networks. Similarity-guided pruning reduces their parameter counts by 39-75% while decreasing mean Dice by less than 0.4 percentage points. To examine token mixers within blocks, we introduce Anatomical Relation Distortion, which quantifies how bypassing a mixer distorts feature relations between labeled anatomical positions. Applied to an unpruned candidate architecture, Centered Kernel Alignment indicates high representational redundancy in intermediate encoder mixers, whereas Anatomical Relation Distortion shows that they contribute little and the bottleneck mixer most. These findings motivate SlimFormer++, which uses Selective Mixing blocks with windowed Mamba mixing in the first encoder stage and only residual channel MLPs in deeper stages. At the bottleneck, Augmented Mixing injects similarity-weighted sparse long-range context into the axial taps of a depthwise convolution at cost linear in voxel count. Evaluated against 12 baselines across eight public CT, MRI, and ultrasound datasets (1,809 cases), SlimFormer++ achieves the highest mean Dice (87.84%) and lowest mean HD95 (6.50 mm) with the fewest parameters (5.98 million). These results support stage-aware token mixing as a design strategy for improving the accuracy-efficiency trade-off in volumetric segmentation. The source code is available at https://github.com/WHHHHHY822/SlimFormerPlusPlus

0.33

Generation-Heavy Hybrid Machine Translation

cs.CL ·2026-10-18 ·Nizar Yahya Habash
Abstract · Dissertation submitted to the Faculty of the Graduate School of the University of Maryland at College Park in partial fulfillment of the req…

The state of the art techniques in Machine Translation (MT) require large amounts of symmetric resources from source and target languages. This is true regardless of whether the approach is Transfer or Interlingua, Symbolic or Statistical. Symmetry within these approaches is necessary to ensure quality, robustness and retargetability. In reality, such symmetry, whether it is in terms of structural transfer lexicons, interlingual dictionaries or parallel corpora, is a major bottleneck in developing any MT system. This dissertation presents an approach to MT that addresses the lack of symmetry by exploiting symbolic and statistical target language resources in source-poor/target-rich language pairs. This approach is called Generation-Heavy Hybrid Machine Translation (GHMT). Expected source language resources include a syntactic parser and a simple one-to-many translation dictionary. No transfer rules or complex interlingual representations are used. Rich target language symbolic resources are used to overgenerate multiple structural variations from a target-glossed syntactic dependency representation of source language sentences. Statistical target-language resource are then used to select amongst the overgenerated translations. The source-target asymmetry of systems developed in this approach makes them more easily retargetable to new source languages.

0.33

What Can't Jev Decide? Probing the Limits of One-Pass Decision Models

cs.CL cs.AI ·2026-10-18 ·Wang Yang, Xinpeng Li, Yanyan Zhang, Chuang Ma, Shouren Wang +6 more
benchmarkhot direction
Abstract

Jev and its open replicas choose among the options a caller supplies with a single forward pass, fast enough to decide inside an agent loop. We ask what such a model cannot decide, and find a simple limit: Jev cannot make a decision that an LLM makes only by writing out intermediate steps. On a benchmark that varies one factor at a time, Jev finds and selects records without error up to 20k tokens and 64 options, but once the answer depends on two or more updates of one user, accuracy falls, to 53-80% at sixteen updates, because similar records interfere. Given a stale summary of the state, Jev follows it. On real decision items graded by how many thinking tokens a reference LLM needs, Jev answers 97.5% of the items that one forward pass of that LLM solves, but only 74% of the others. Jev also does not know its limit: its confidence flags the errors it makes itself but not those it inherits from a stale input, and neither an explicit nor a trained hand-off option catches the latter. The open models we evaluate share these limits. Checking the state such models rely on requires a model that reasons.

0.32

Probing and Mitigating Hallucinations in Speech-augmented Language Models for Automatic Speech Recognition via Small Language Models

eess.AS ·2026-10-18 ·Bi-Cheng Yan, Jhih-Rong Guo, Fu-An Chao, Berlin Chen
published
Abstract · 5 pages, 5 figures, accepted by INTERSPEECH 2026

The widespread success of large language models in natural language processing has spurred the development of speech-augmented language models (SLMs) to extend their reasoning capabilities from textual inputs to spoken ones. Despite impressive performance, SLMs remain susceptible to hallucinations, posing challenges in practical utility. To this end, we first probe hallucinations in SLMs for automatic speech recognition (ASR) via the lens of causal mediation and behavioral analysis. To mitigate fabricated ASR hypotheses, we propose AudioSLM, a compact yet effective ASR framework built upon a small language model, which leverages alignment cues and cross-attention layers to synergistically bolster acoustic modeling and cross-modal interactions. Experiments on LibriSpeech reveal that hallucinations arise from multi-head self-attention modules with an excessive attention bias towards textual tokens. AudioSLM considerably eases ASR hallucinations and outperforms some LLM-based ASR models.

0.32

Sharp Degree Bounds for Putinar Certificates in Terms of the Positivity Margin via Analytic Sums of Squares

cs.CL ·2026-10-18 ·Yuxiang Zhao, Ting Gan
theory
Abstract

We establish sharp degree bounds for Putinar sum-of-squares (SOS) certificates in terms of the positivity margin $δ$. For any fixed Archimedean quadratic module with prescribed generators and any fixed degree bound on the objective, normalized polynomial objectives admit a uniform certificate-degree upper bound of $O(δ^{-1/2}\log^2(2/δ))$. The bound is sharp: we construct a matching lower-bound example in three variables. We further show that the number of inequality generators affects the optimal degree bound. For quadratic modules with at most two inequality generators, the bound improves to $O(δ^{-1/2}\log(2/δ))$, which is also sharp. This improvement is obtained by directly converting two-generator Stengle refutations into certificates for the original quadratic module. We also establish sharp degree bounds for trigonometric SOS certificates on tori. In particular, for the full torus without additional constraints, the optimal worst-case order improves to $δ^{-1/4}$, whereas additional equality constraints can lead the order to $δ^{-1/2}\log(2/δ)$. These results also yield sharp worst-case convergence rates for the Lasserre hierarchy.

0.32

PLANE: Plug-and-Play Fine-Grained Visual Experts for Hallucination Mitigation in Large Vision-Language Models

cs.CV cs.AIcs.CL ·2026-10-18 ·Jianghao Yin, Lang Yu, Qianxi Yu, Qin Chen, Jie Zhou +2 more
top venuebenchmarkdataset
Abstract · Accepted by EMNLP2026 Findings

Large Vision-Language Models (LVLMs) commonly suffer from hallucinations, which appear as fabricating non-existent objects or outputting incorrect attributes and spatial relationships in images. Previous works empirically indicate that CLIP-style vision encoders tend to focus on primary objects while ignoring image details, which is one of the causes of hallucinations. To address this issue, we propose plug-and-play fine-grained visual experts (PLANE) comprising two components: LoRA visual experts operating on CLIP for multi-aspect image encoding, and an instruction-aware gate network for adaptive image feature fusion. During training, we construct a fine-grained image-caption pairs dataset, where images are described from three aspects: objects, attributes, and relations between objects. Image-caption pairs from different aspects are used to train the three fine-grained LoRA experts respectively. Aiming at comprehensively evaluate our method's capabilities, we manually annotate a fine-grained hallucination evaluation benchmark for LVLMs which is GPT4-independent during evaluation phase (GIFT). We conduct experiments both on popular hallucination evaluation benchmarks and GIFT. The results show that our method effectively reduces hallucinations across models of different types and sizes while introducing minimal inference latency. Compared to advanced baselines, our model also demonstrates superior performance.

0.31

Morpheus: a fully local, defense-in-depth de-identification engine for Greek clinical text - a benchmark evaluation

cs.CL ·2026-10-18 ·Filippos-Paschalis Rorris, Nikoleta Stanitsa, Tatiana Rorri, Foteini Rorri, Constantine N. Antonopoulos
benchmark
Abstract · 24 pages, 2 figures, 3 tables; supplementary material appended. Benchmark corpus: doi:10.5281/zenodo.22824894

Background and Objective: Clinical free text cannot leave the hospital without de-identification, yet no de-identification system or benchmark exists for Greek clinical text. We evaluated Morpheus, a fully local de-identification pipeline for Greek clinical documents, on 500 synthetic discharge summaries. Methods: The notes (300 cardiac surgery, 100 cardiology, 100 nephrology) were generated by a frontier large language model (Gemini 3.1 Pro) constrained to authentic Greek hospital document layouts. Notes were pre-annotated at generation; two clinicians then independently reviewed and corrected every note, blinded to each other, with disagreements arbitrated. Primary outcomes were recall over must-redact identifiers and over-redaction of clinical content, measured over must-survive probes. Secondary outcomes: residual leaks flagged by an independent egress-review layer, processing time and reproducibility. Results: Inter-annotator agreement was F1 = 0.998. The de-identification rate was 96.59% (7,666/7,937 identifier instances; 95% CI 96.16-96.96): 98.49% for direct and 93.42% for quasi-identifiers. No patient's own complete identifier survived in any note; of 271 residual instances, 199 (73.4%) concerned institutional identity, chiefly the hospital letterhead. The over-redaction rate was 2.82% (190/6,747 probes; 95% CI 2.45-3.24). The egress guard flagged 2 of 271 residuals, both complete telephone numbers. Two independent runs gave identical verdicts. Conclusions: A fully local, deterministic pipeline removed 98.5% of direct patient identifiers from Greek clinical text while altering under 3% of protected clinical content; most residuals concerned the institution, not the patient. To our knowledge this is the first benchmark evaluation of clinical-text de-identification for Greek and the first quantified evaluation of an independent egress-guard layer.

0.31

Where Are Language-Specific Features Represented in Multilingual LLMs? Everywhere

cs.CL ·2026-10-18 ·Mikaela Martins, Jonathan Dunn
Abstract

Previous work has argued that multilingual LLMs encode language-specific features most strongly in their initial and final layers, implying that intermediate layers support language-agnostic representations. This paper reconsiders this finding by performing four different experiments: three that undertake a fine-grained analysis of representations of Portuguese and Spanish, two closely related languages, and one large-scale language identification experiment that covers over 200 languages. The results of these experiments challenge the strict separation between language-specific and language-agnostic layers in multilingual LLMs, using Llama 3.2 1B as a test case. By analyzing and manipulating contextual sentence embeddings, we demonstrate that language identity remains strongly recoverable throughout the model rather than being restricted to initial and final layers. Moreover, we find that the model tends to classify ambiguous strings as closer to Spanish than to Portuguese, revealing a skew in model behavior that differs from human judgments.

0.31

Aligning Cultural Behaviors in Arabic Language Models via Preferences

cs.CL cs.AI ·2026-10-18 ·Junhong Liang, Daniil Orel, Saeed Almheiri, Bilal Elbouardi, Younes Samih +9 more
top venuebenchmark
Abstract · Accepted to EMNLP 2026

Arabic LLMs have improved rapidly, but aligning them with country-specific cultural behavioral preferences remains underexplored. Previous work has focused primarily on instruction-tuning for factual knowledge, overlooking nuanced cultural behaviors. We propose ArabPref, which to our knowledge is the first large-scale pan-Arab preference data set for the alignment of cultural behavior conditioned on the country. Our contribution includes two resources in English and Modern Standard Arabic: a culturally grounded preference dataset for training and evaluation, and a multiple-choice benchmark designed to assess culturally aligned behavior across nations. All test data are validated by native speakers to ensure authenticity. To optimize cultural alignment, we experiment with post-training methods across multilingual and Arabic-centric language models, evaluating performance on both generation tasks and our cultural multiple-choice benchmark. By releasing an evaluation benchmark, we provide a foundation for systematically studying and improving culturally aligned behavior in language models.

0.31

A Thousand Faces: A Persuasion Benchmark with Experience-Grounded Persuadees Informed by Psychology and Pragmatics

cs.CL ·2026-10-18 ·Ziqing Zhuang, Dingyi Zhang, Deyu Zhou
benchmarkcode/data
Abstract · 45 pages. Dataset: https://huggingface.co/datasets/Slandes/ATF-Persu

People can express the same beliefs and desires yet need different reasons to change their minds. What makes an argument relevant depends on how those views arose: experiences shape what a person finds credible, what matters to them, and what could prompt reconsideration. These personal grounds may also be expressed indirectly, through contextual clues in conversation. Benchmarks that specify beliefs and desires without these connections can reward arguments that match a description while overlooking whether they address the person's reasons. We introduce A Thousand Faces (ATF-Persu), with 1,000 personas across 200 topics and nine domains. Drawing on personality and attitude theories and conversational implicature, we construct five personas per factual context, linking experiences to judgments, expression, and grounds for revision. Evaluations show lower persuasion success than three comparison benchmarks under their respective protocols. Complete backgrounds yield broader use of supplied personal grounds and higher persona faithfulness than simplified specifications; blinded reviewers also rate ATF-Persu above CToMPersu on identifiable personal reasons, their influence on responses, and contextual expression. The dataset is publicly available at https://huggingface.co/datasets/Slandes/ATF-Persu.

0.31

MCSplat: Multi-View Photometric and Geometric Consistent Feed-Forward Gaussian Splatting for Driving Scenes

cs.CV ·2026-10-18 ·Yifei Wang, Wei Tian, Wenhua Bao, Jia Zhou, Hanshi Wang +8 more
top venue
Abstract · 24 pages, 14 figures, 6 tables; Accepted to NeurIPS 2026

Driving-scene reconstruction via feed-forward Gaussian splatting from sparse multi-view images remains challenging due to two factors. First, cross-camera photometric inconsistency, caused by variations in exposure, ISP, and illumination, can be incorrectly absorbed into Gaussian geometry. Second, unreliable local geometric supervision arises from normal-based constraints near depth discontinuities. To address these issues, we present \textbf{MCSplat}, a multi-view consistent feed-forward Gaussian splatting framework for driving scenes. MCSplat constructs a canonical appearance space from depth-guided multi-camera correspondences to decouple geometry learning from camera-specific appearance. A multi-scale Bilateral Grid Transformer restores raw camera appearance from canonical renderings, enabling photometric supervision without contaminating Gaussian geometry. To improve local geometric reliability, we estimate Fisher-information-based uncertainty from photometric sensitivity and use it to adaptively weight surface-oriented self-supervision. Extensive experiments demonstrate improved reconstruction quality and generalization across driving datasets. Geometry-oriented evaluations further reveal enhanced cross-view depth consistency and local surface-normal agreement, while qualitative results show fewer geometric artifacts and more coherent structures across overlapping views.

0.30

EmotiQ: A Quantum Residual Specialist Framework for Text Emotion Classification

cs.CL ·2026-10-18 ·Carlton Kigonya, Tanaka Chirara, Yunting Yin
published
Abstract · Accepted to IEEE ICTAI 2026

Emotion classification from text is challenging because many affective states are distinguished by subtle linguistic signals and semantically overlapping boundaries. We investigate whether compact quantum modules can serve as targeted residual specialists that assist traditional classifiers on difficult emotion decisions. We introduce EmotiQ, a residual specialist framework for text emotion classification using a frozen transformer classifier as the primary decision maker, identifies challenging emotion boundaries from validation behavior, and trains boundary specialists, including quantum encodings, data re-uploading specialists, prototype and fidelity specialists, and kernel-based specialists. Specialists are permitted to intervene only when they satisfy validation-based admission criteria. We evaluate the framework across six text emotion classification datasets and establish traditional model baselines. Experimental results show that the admitted EmotiQ policy improves mean macro-F1 metrics over the baseline, and achieves positive gains on five of six datasets. Among the quantum components, ZZ interaction feature maps are the most consistently effective, yielding a mean macro-F1 improvement of +0.0030 and positive gains on four datasets. These findings suggest that current quantum NLP methods can be deployed as residual specialists that provide localized improvements under strong classical supervision.

0.30

When Does Better Acceptance Pay? A Controlled Study of Speculative Decoding

cs.CL ·2026-10-18 ·Yongchao Long, Chunrui Zheng, Yuxuan Zhou, Guangkun Nie, Tian-Chi Ma +8 more
code/data
Abstract · 24 pages, 11 figures. Project page: https://github.com/ernie-research/DCore

The throughput gains of speculative decoding depend on the trade-off between acceptance length and execution latency. We ask when additional acceptance is worth the cost of obtaining it. We establish a comparison protocol spanning training, inference, and evaluation, remeasuring released checkpoints and retraining six drafter recipes on shared target-generated data across ten benchmarks and two target models. Alignment changes the picture: gaps that appear large in published comparisons shrink substantially under shared training and inference conditions. This narrowing motivates two deeper questions: how is acceptance improved at different costs, and when do these gains improve throughput? We find that acceptance can be improved through nearly free objective and context changes or more expensive token-to-token conditioning, which pays off most clearly under sampled drafting. DCore illustrates the former route as a strong baseline with little additional measured drafting overhead. The payoff is conditional: task uncertainty and serving conditions affect acceptance returns, execution costs, and the preferred drafter and budget. These findings inform future drafter design and deployment choices.

0.30

Stable Signer V3: 4D Deformable Primitive Splatting Is All You Need for Sign Language Production

cs.CV cs.AIcs.CLcs.GR ·2026-10-18 ·Sen Fang, Liwei Che, Dimitris N. Metaxas
code/data
Abstract · 12 figures, 10 tables. Project Page at: https://signerx.github.io/Stable-Signer-V3

We present Stable Signer V3, an interactive high-resolution sign language avatar system built on 4D Deformable Primitive Splatting. It addresses three limitations of existing avatars: restricted facial expression, inaccurate identity-specific body deformation, and loss of fine appearance details. Shared Face Primitive Splatting adds transferable superquadric primitives for richer expression control; Test-Time Body Primitive Splatting learns an identity-specific virtual muscle layer for more faithful motion; and Test-Time Tunneling Super-Resolution uses TunnelingNet to recover identity-specific high-resolution details. Both test-time modules are optimized once per signer identity; inference then runs entirely from the stored identity parameters. Together, they reduce facial geometry error by 71.4%, body-boundary error by 35.1% across ten identities, and high-resolution RGB reconstruction error by 24.2% on held-out evaluations, while retaining interactive streaming.

0.30

Does Expert Supervision Improve Text-to-Speech? An Empirical Study of Preference Optimization

cs.SD eess.AS ·2026-10-18 ·Minh Duc Vu, Junchuan Zhao, Ye Wang
Abstract · 19 pages, 10 figures, 11 tables

Preference optimization promises to improve text-to-speech (TTS) systems by learning which speech outputs are better. External expert speech seems a natural source of preferred responses, especially when student generations contain errors. Yet demonstration quality does not establish whether the student can benefit from that supervision, or whether better target fitting will produce better speech. We examine this gap across two TTS students using real and synthetic expert speech, comparing preference objectives and direct versus bridge supervision while tracking fixed-sequence likelihoods and generation metrics. With expert supervision held fixed, objectives differ markedly in their sensitivity to preference-score scaling and the resulting generation quality. Likelihood diagnostics characterize these responses, but cannot establish their benefit: expert likelihood can increase while intelligibility worsens, or decrease while it improves. Bridge supervision highlights a distinction between mitigating degradation and improving the student: it can recover performance lost under direct training without consistently surpassing the pretrained base. Route benefits vary across objectives, students, and expert sources; neither route consistently dominates. Together, these findings distinguish demonstration quality, target fitting, and generated-speech quality as complementary criteria for assessing expert supervision.

0.30

From Label Priors to Task Evidence: Efficient Long-Video Frame Selection via Bayesian GRPO

cs.CV ·2026-10-18 ·Shuochen Chang, Bingjie Gao, Qingyang Liu, Qianli Ma, Yibo Miao +5 more
code/data
Abstract

Multimodal Large Language Models (MLLMs) face significant scalability challenges in long-video understanding due to quadratic attention complexity and context window constraints. Existing retrieval-based approaches typically rely on supervised learning from static keyframe annotations. However, these label priors often fail to align with the actual information required by downstream MLLMs for complex reasoning, creating a misalignment between the training target and inference needs. To bridge this gap, we propose a probabilistic framework that reformulates frame selection as latent-variable inference. We first initialize a lightweight policy using supervised learning to capture general visual saliency. Then, we treat a frozen MLLM as a task environment and use its answer likelihood as dense task evidence to refine the policy via Bayesian Group Relative Policy Optimization (GRPO). Within this formulation, the negative log-likelihood (NLL) over ground-truth answer tokens provides a principled dense reward, while a prior-matching KL term preserves semantic consistency with the elicited label prior. This process transitions the selection principle from static label priors to task evidence. Extensive experiments on Video-MME, MLVU, and LongVideoBench demonstrate that our selector achieves competitive accuracy with MLLM-based scorers while retaining the inference efficiency of lightweight methods. The selector is trained once with a frozen MLLM environment and transfers directly to multiple downstream MLLMs without retraining. Code is available at github.com/Miraclecsc/Frame_Selection_GRPO

0.29

MT Forcing: Towards Stable and Consistent Multi-Turn Image Editing with Autoregressive Diffusion

cs.CV ·2026-10-18 ·Houyuan Chen, Fanzhou Wang, Xiangyu Fan, Zesong Qiu, Yukun Wei +8 more
code/datacommunity pick
Abstract · Project page: https://houyuanchen111.github.io/mt-forcing.github.io/

Multi-turn image editing aims to modify an image according to a sequence of instructions, enabling more effective interactive visual content creation for users. However, existing methods often exhibit noticeable drift after only a few editing turns, manifesting in two aspects: 1) unstable visual quality, with accumulating artifacts and degraded details; and 2) inconsistent image content, with unintended changes to preserved regions and prior edits. To address these challenges, we propose MT Forcing, a novel training paradigm for multi-turn image editing with unified multimodal models (UMMs) based on autoregressive diffusion. Specifically, we propose Single-Turn Distribution Matching Distillation (Single-Turn DMD), which supervises student self-rollouts with a teacher that applies cumulative edits to the source image in a single turn, providing visual-quality guidance without recursive error accumulation. Additionally, we introduce Teacher-Forcing Guidance, which supervises history-dependent transitions on reference trajectories during warmup and as a regularizer during distillation, encouraging cross-turn content consistency beyond what per-turn distribution matching alone can ensure. Extensive experiments demonstrate that our approach significantly improves the multi-turn image editing capabilities of UMMs, yielding more stable visual quality and stronger cross-turn consistency.

0.29

Visual Sufficiency Landscapes: Understanding Query-Conditioned Reliability in Visual Token Pruning

cs.CV ·2026-10-18 ·Zhenghan Li, Kexin Sun, Wei Zheng, Peng Hu, Xiao Wei +1 more
Abstract

Different visual token pruning strategies retain different subsets while maintaining similar task performance. This motivates a combinatorial question: how many visual subsets can preserve model behavior, and how does their abundance depend on the query? We introduce visual sufficiency landscapes to characterize this abundance. At a fixed layer and budget, visual sufficiency probability is the fraction of uniformly sampled subsets that jointly preserve the dense model's answer, reference-answer likelihood, and output distribution. Independent sampling replicates separate input variation from estimation noise. Across Qwen2.5-VL-7B, InternVL2-8B, and LLaVA-v1.5-7B, differences between inputs on GQA and TextVQA exceed replication error, and different questions about the same GQA image produce distinct landscapes. Sufficiency probability also improves failure prediction for full pruning methods. On LLaVA, adding this probability increases AUC by 0.1151 for FastV, 0.0683 for VisionZip, and 0.1401 for SparseVLM, while reducing log loss and Brier score. Several architecture-adapted configurations on the other two models reproduce these gains. Under controlled selection rules, predictors trained on the original images reduce log loss on new images by 9.0%, 13.0%, and 14.1%, respectively. These results establish successful-subset abundance as a measurable dimension of local prunability and connect it to the identification of high-risk pruning conditions.

0.29

MixiSpeech: Progressive State Distillation for One-Step Vietnamese TTS

cs.SD cs.MMeess.AS ·2026-10-18 ·Tran Quoc Bao Le, Binh Mai, Hung Dinh, Lam Nguyen, Cong Tran
publishedcode/data
Abstract · Accepted at SOICT2026

Conditional-flow text-to-speech (TTS) models can achieve high synthesis quality with relatively few sampling steps, but collapsing the remaining trajectory to a single update can still cause substantial quality degradation. We propose MixiSpeech, a progressive state distillation framework that converts a multi-step conditional-flow acoustic model into a one-step generator while preserving its architecture and speaker conditioning. Rather than training the one-step model directly, MixiSpeech progressively reduces the global sampling budget through a 10-4-2-1 curriculum, with each student trained to match local teacher-integration endpoints using a masked state-matching objective. We instantiate the framework on a 710-speaker Vietnamese Matcha-TTS model trained from scratch on PhoAudiobook. On 100 shared test text-speaker pairs, direct one-step inference substantially degrades synthesis quality, obtaining 1.6846 UTMOS22-Strong and 6.21% WER compared with 2.2923 and 4.65% for the 10-step teacher. In contrast, the distilled one-step model achieves 2.2678 UTMOS22-Strong, 4.63% WER, and 2.83% CER, recovering most of the quality lost at the one-step operating point while requiring only a single acoustic vector-field evaluation. End-to-end inference time decreases from 0.0662 s for the 10-step teacher to 0.0312 s for MixiSpeech, corresponding to a 2.12 measured speedup. These results show that progressive state distillation can make one-step conditional-flow TTS practical without relying on a large multi-step sampling budget. Project page: https://mixispeech.solanai.us/

0.29

Reliability-Preserving Compression for Deployable Medical Image Segmentation

eess.IV cs.AIcs.CVcs.LG+1 ·2026-10-18 ·Sagar Kumar, Shailja Sharma, Tirthajit Baruah, Punit Rathore
Abstract

Post-training compression of pretrained medical segmentation models commonly uses quantization, pruning, or distillation to reduce model size and computational cost. Existing medical segmentation quantization studies mainly evaluate Dice, model size, memory, and latency, but this is incomplete for safety-critical use because compressed models should also preserve uncertainty signals that indicate unreliable predictions. We study this problem in compressed MRI/CT segmentation and formalize why Dice can remain stable while uncertainty changes. Motivated by this failure mode, we propose score-guided aggregate-validated mixed precision (SGAV-MP), a post-training compression framework that selects layer precisions under reliability constraints. SGAV-MP ranks quantization actions with a compact benefit-risk score and accepts them only after full-model reliability validation on a calibration set. Across five 2D backbones and three 3D benchmarks, SGAV-MP preserves FP32-level Dice and high uncertainty-map agreement, achieving $3.59\times$--$7.8\times$ compression in 2D and $4.73\times$--$7.82\times$ in 3D. We also report backend-supported TensorRT deployment measurements separately to distinguish runtime execution from reliability-preserving storage compression.

0.28

A Multi-granularity Interaction Network for Relational Triple Extraction

cs.CL ·2026-10-18 ·Xiaocheng Luo, Yanping Chen, Zaifu Zhan
code/data
Abstract

Relational triple extraction identifies entity relations from raw sentences without depending on entity type information. The performance of traditional methods is seriously degenerated by ``unseen'' entities, because these methods can't build contextual dependencies relevant ``unseen'' entities. This paper addresses this challenge by proposing a multi-granularity interaction network that leverages word-word, entity-relation, and head-tail interactions to capture deep semantic dependencies. By incorporating multiple iterations, the multi-granularity interaction network is effective to refine the representations corresponding to head and tail entities. It enhances the global semantic features associated with relational triples, aiding in the remediation of triples that may have been overlooked or omitted in earlier steps. Strong and consistent performance is achieved by our method on two rearranged datasets, with a significant improvement in identifying relational triples, including those involving ``unseen'' named entities. For example, our model improves F1-score by 3.1%/20.5% over previous approaches on the NYT-R/WebNLG-R datasets in the Unseen metric. Code: https://github.com/xiaocen-luo/Interaction-Network.

0.27

LateMatch: Breaking the Interaction Bottleneck via Cacheable Matching for Efficient Visual Localization

cs.CV cs.RO ·2026-10-18 ·Mohammad Omama, Aeyan Ashraf, Sandeep Chinchali
Abstract

Visual localization is critical for embodied agents, but deploying state-of-the-art deep feature matchers on edge devices is fundamentally limited by their massive computational cost. Standard deep matchers tightly interleave intra-image context building and inter-image matching at every layer, creating a severe interaction bottleneck. Localizing a single query against K database images forces the system to execute the full network K independent times with no capacity for caching. To break this bottleneck, we advocate for a paradigm shift toward cacheable image matching. We realize this paradigm with LateMatch, a versatile approach that restructures existing matchers by deferring inter-image interaction to their final layers. By isolating intra-image processing into offline-cacheable Context Blocks, only a small set of Interaction Blocks must be executed online. To further optimize LateMatch, we introduce E-SToRE (Entropy-guided Spatial Token Reduction and Expansion). E-SToRE leverages the decreasing Shannon entropy of the late Interaction Blocks to dynamically reduce token counts just prior to cross-image interaction. Evaluated across HPatches, ScanNet, MegaDepth, ETH-3D, and Aachen, LateMatch reduces effective computational cost by over 2x compared to standard deep matchers, while E-SToRE yields an additional 10% to 14% savings with negligible degradation in accuracy.

0.27

Atoms, Not Knowledge: What Production-Scale Addressed-Memory Writes Deposit

cs.CL ·2026-10-18 ·Wei Gao, Cheng Tian, Duan Yang, Jianqing Yin
Abstract

Deterministic hash-addressed conditional memory permits local post-hoc knowledge injection. Extending arXiv:2610.21553's bounded-feasibility audit rejecting under-specified generality claims with repaired probes and training-dose/capacity axes, we give, to our knowledge, the first production-scale write-content verdict. We test 124M-7B self-trained/publicly pretrained beds, Qwen3.8-Flash-Next-FP8 (176.95B; 320,001,536 FP8 rows), and DeepSeek-V4.1-Flash. Write coverage requires a sufficiently trained backbone prior, full-width readout, trainable gates; undertrained-bed intervention confirms a threshold, not a slope. At fixed budget, coverage falls from 98% (50 facts) to 12.7% (1,000); table size is not binding. Two-hop composition does not exceed the independent single-hop baseline (26% observed vs 25% expected; 15% lexical-distortion floor). Synergy remains absent across dose/load in two independent bed families. A 28-percentage-point in-context gap (given vs recalled intermediates) localizes failure to retrieval-for-reasoning, not gate access. Writes are independently retrievable, precisely addressable, reversible atoms, not composable into reasoning. A layered protocol, the first, to our knowledge, to cover compositionality in behavioral memory evaluation, separates baseline, synergy and distortion floor. Any addressed/key-value memory must report coverage, synergy ratio and distortion rate before deployment claims. Composition must be system-orchestrated; multi-constraint queries should explicitly chain written atoms. In-context controls show given intermediate entities are exploited at twice the implicit rate. Production load fits exponential interference ($R^2=0.98$); its mechanism, between budget dilution and write interference, remains open. All numbers are recomputed from archived raw logs/JSON; a fresh 52-target production rerun gives 52/52, exactly matching archived baselines.

0.27

GroundedVDR: Grounding Video-to-Web Evidence Dependencies in Video Deep Research

cs.CV ·2026-10-18 ·Xiyuan Liang, Yu Wang, Zuhao Yang, Ziting Wang, Duo Li +5 more
benchmarkcode/data
Abstract

Video large multimodal models have improved at identifying cross-frame visual evidence and conducting open-web retrieval, as reflected in recent video deep research (VDR) benchmarks. However, existing VDR benchmarks primarily evaluate tool-use trajectories and final responses without verifying whether a model actually uses the temporal video information or the retrieved web evidence. To fill this gap, we introduce \method{}, an evidence-grounded benchmark with three features. First, it combines human review with restricted-access probing across multiple models to exclude questions answerable from parametric knowledge or a single information source. Second, each question carries structured video-to-web evidence annotations that link the relevant video segments to their supporting web evidence. Third, its evaluation protocol separately scores video grounding, web evidence grounding, and final-answer correctness. Across eight multimodal agent systems, the best reaches 76.0\% answer accuracy, yet none recovers the complete required web evidence for more than 8\% of questions. Our benchmark is publicly available at \url{https://huggingface.co/datasets/hh721/vdr_temp}.

0.27

Only Means Only: what a harmonization benchmark score averages over

cs.CV ·2026-10-18 ·Imad Buljić, Delaida Muminović
benchmarkunusual
Abstract · 18 pages including references, 3 figures in the paper and 1 in the supplement. The supplement is included as an ancillary file and published…

Image harmonization methods adjust the colour and lighting of an object pasted into a photograph so that it looks native to its new scene, and they are ranked by average error on one benchmark whose test images are real photographs with one object recoloured. One photograph can yield several such test images, and the benchmark promises that no photograph appears in both the training and the test split. We find that 1,399 photographs do, and the composites built on them are a quarter of the test set. Duplicate checks over file names or bytes miss this, because every composite is a different file. On that quarter the three methods we score end to end read two to five times lower error than on the rest, but most of the gap is composition rather than the sharing: the shared composites come almost entirely from one of the benchmark's four source collections, and inside that collection the gap falls to about one and a half times. An official score is therefore an average over two populations, mixed in a proportion the benchmark does not report. We release a test list of 5,581 composites whose construction identities do not appear in training, with four confirmed near-duplicate pairs disclosed rather than removed, adoptable without retraining at the price of one rescore per checkpoint. We leave open whether the sharing itself changes what a model learns.

0.26

UBD-Bench: Benchmarking and Analyzing Unimodal Branch Degradation in Multimodal Perception for Autonomous Driving

cs.CV ·2026-10-18 ·Ang Liu, Xiang Gu, Ruikai Li, Qiyue Wang, Yizhe Fan +8 more
benchmark
Abstract · 11 pages, 5 figures

Multimodal perception improves autonomous-driving performance by combining complementary sensor information. However, stronger multimodal predictions do not necessarily imply stronger representations in each modality-specific branch. Such branch-level weaknesses can remain hidden in conventional output-level evaluation. We introduce UBD-Bench, a systematic benchmark for studying unimodal branch degradation (UBD) in multimodal perception. UBD-Bench uses a Paired Frozen-Branch Probe to compare the task-relevant representation quality of the same modality after independent unimodal training and multimodal joint training under matched decoding conditions. Across four autonomous-driving perception tasks and 10 models, we find that Camera branches consistently degrade after multimodal joint training, whereas LiDAR branches exhibit weaker and more model-dependent changes, revealing a pronounced modality asymmetry. Hierarchical probing further shows that Camera-branch degradation extends from the visual backbone through the Camera-to-BEV pathway and is most severe at the pre-fusion Camera BEV representation. Guided by this diagnosis, we introduce unimodal-to-multimodal distillation at the degraded representation. This intervention consistently improves both Camera-branch Probe performance and full-model performance across all four tasks. These results show that multimodal gains can coexist with degraded unimodal representations, highlighting branch-level representation quality as an important dimension for evaluating and improving multimodal perception.

0.26

MoWE: Spatially Adaptive Mixture-of-Wavelet-Experts Pooling for Ultrasound Lesion Segmentation

cs.CV ·2026-10-18 ·Reaz Hassan Joarder, Sidratul Tanzila Tasmi, Alif Arshad Bakshi, Md. Bakhtiar Hasan, Md Hasanul Kabir
published
Abstract · Accepted to WACV 2027

We present Mixture-of-Wavelet-Experts (MoWE), a spatially adaptive routing mechanism that replaces max-pooling and strided convolutions in U-Net architectures for ultrasound lesion segmentation. Standard pooling operations suppress high-frequency spatial details, including potential boundary cues in speckled regions. Recent frequency-domain methods preserve some of these details, but a single wavelet family applies the same spectral bias across the image. MoWE instead uses a context-aware router to assign local regions to fixed wavelet families according to local texture and global feature context. The decoder reuses the encoder's spatial routing maps during the inverse transform. Learnable channel-wise gates rescale the high-frequency subbands before reconstruction. We evaluate MoWE with U-Net on four ultrasound lesion benchmarks (BUSI, TN3K, UDIAT, BUS-UCLM) and one cardiac dataset (CAMUS), and with nnU-Net on BUSI. On BUSI, MoWE improves U-Net Dice from 72.59\% to 74.55\% and reduces HD95 from 47.01 to 43.49\,px. With nnU-Net, MoWE improves Dice from 80.18\% to 81.97\% and reduces ASSD from 20.07 to 17.25\,px. On U-Net, MoWE reduces trainable parameters by 6.4\%.

0.26

Image Dehazing as In-Context Tabular Inference

cs.CV ·2026-10-18 ·Le-Anh Tran
benchmark
Abstract · 10 pages, 6 figures, 8 tables

Image dehazing is traditionally addressed through handcrafted physical priors or task-specific deep learning models, often requiring carefully designed architectures or extensive model training. Recent advances in tabular foundation models (TFMs), however, offer an alternative learning paradigm in which continuous relationships can be inferred directly from contextual examples without task-specific model training. Motivated by this capability, this paper investigates whether single-image dehazing can be effectively reformulated as an in-context tabular regression problem and introduces DehazeTab, a framework that estimates atmospheric transmission from structured image-patch representations. Specifically, overlapping patches extracted from hazy images are transformed into tabular samples using handcrafted haze-related descriptors, pretrained visual embeddings, or their hybrid combinations, while corresponding transmission patches are summarized as scalar regression targets. A pretrained TFM then performs in-context regression using labeled reference patches to predict transmission values for unseen images without any further fine-tuning. These predictions are spatially aggregated into a transmission map, refined through guided filtering, and incorporated into the atmospheric scattering model to recover haze-free images. Experiments on benchmark datasets demonstrate competitive restoration performance and reveal that physically motivated handcrafted features can outperform pretrained visual embeddings, while hybrid representations offer complementary benefits. Further analyses examine the effects of feature representations, patch selection, and contextual reference configurations on restoration quality. Beyond image dehazing, the findings highlight the potential of TFMs as an alternative paradigm for addressing low-level vision tasks through structured and in-context inference.

0.26

Dual-Slot: From Serial Loops to Parallel Slots in Language Models

cs.CL ·2026-10-18 ·Yiqin Hao, Zitong Wang, Boyi Zeng, He Li, Shixiang Song +5 more
Abstract · 17 pages, 6 figures

Looped language models increase computational depth by repeatedly applying shared layers, offering a way to improve model quality without adding parameters. However, sequential dependencies between loop iterations increase autoregressive decoding latency. We introduce Dual-Slot, an architecture that reorganizes serial looped computation into parallel latent and prediction slots. Rather than waiting for the current token's latent computation to finish, the prediction slot combines the previous token's deep latent state with the current representation, while the latent slot produces a state for the next decoding step. This shifts the latent dependency across adjacent tokens, replacing two sequential module invocations with a single paired invocation while preserving standard autoregressive generation. To preserve token-level parallelism during training, we use a small number of Jacobi iterations to refine latent states across the sequence in parallel. Experiments on Qwen-style dense and mixture-of-experts (MoE) models show consistent improvements in language modeling and average downstream accuracy. On the 1B dense model, Dual-Slot reduces Pile validation perplexity from 7.25 to 7.00 and improves average five-shot accuracy across nine task configurations by 2.26 percentage points. After supervised fine-tuning, it improves zero-shot GSM8K accuracy by 5.83 and 3.49 percentage points for the 1B dense and A0.6B MoE models, respectively. Across the evaluated batch sizes, Dual-Slot retains 93.1-96.5% of vanilla decoding throughput for the dense model and 98.1-99.6% for the MoE model.

0.26

SENRI: Understanding the Space-Event Nexus in Real-Life Videos

cs.CV ·2026-10-18 ·Kehan Lan, Kaining Ying, Henghui Ding, Kaipeng Zhang
code/data
Abstract · Project Page: https://senri-video.github.io/

Human memories are organized in a way that connects events with their corresponding spatial contexts. An AI assistant should also be able to memorize and recall the nexus of events and spaces. Thus we introduce SENRI, a task that studies this space--event connection in long, continuous real-life egocentric video. It contains six task types in two directions: Event$\rightarrow$Space, which recovers spatial knowledge from event cues, and Space$\rightarrow$Event, which retrieves events from regions and trajectories. To evaluate this, we build SENRI-Bench, containing 500 manually annotated multiple-choice questions over 10 uninterrupted self-recorded videos with around 25.3 hours of content. Also, we evaluate three typical long video understanding paradigms: direct perception, evidence seeking, and persistent memory. The task remains challenging across all three paradigms, highlighting substantial room for improvement. Then we show two representative failure modes of existing memory methods and address these evidence gaps with SENRI-Agent, a controller that uses Target-Guided Visual Check (TGVC) to propose local visual check tasks for parallel verifiers and integrate the reports into a working memory to help with reasoning. Built upon existing memory systems, SENRI-Agent improves the performance of existing baselines without modifying their stored memories. We hope SENRI will provide valuable insights into how real-life egocentric understanding should be evaluated, reveal the gaps in existing approaches, and offer a possible direction for improvement.

0.26

MultiAltitudeSR: Multi-Altitude Agricultural UAV Image Super-Resolution via Physical Ground Sampling Distance Conditioning

cs.CV ·2026-10-18 ·Xintong Jiang, Yu Tian, Yixue Liu, Valerio Hoyos-Villegas, Viacheslav Adamchuk +1 more
Abstract · 27 pages, 12 figures

Unmanned aerial vehicles (UAVs) enable high-throughput platforms for crop monitoring but face a fundamental trade-off between flight altitude and image spatial resolution. Many existing super-resolution (SR) methods for UAV imagery rely on oversimplified synthetic degradation models and are restricted to fixed integer upsampling scales, limiting their ability to handle real-world degradations and varying flight altitudes. This study proposes MultiAltitudeSR, a framework that conditions feature modulation and reconstruction on source and target ground sampling distances (GSDs). MultiAltitudeSR trained jointly on real UAV images at altitudes of 10 to 70 m above ground level supports all 21 source-to-target altitude pairs without pair-specific fine-tuning, covering scale factors ranging from 1.18$\times$ to 6.03$\times$. The framework can be extended to additional source-target GSD combinations by incorporating corresponding image pairs into training. On the 20-to-10 m pair, the model achieved competitive perceptual performance compared with pair-specific baselines, with LPIPS of 0.2236, DISTS of 0.2431, and NIQE of 8.9020. Experiments using the same architecture demonstrated the value of acquiring real-world multi-altitude UAV image pairs for SR training, with real-pair training achieving better average reconstruction performance across the 21 altitude pairs than training with the tested synthetic degradation pipeline. Downstream zero-shot segmentation case studies illustrated task-dependent benefits of super-resolved imagery for crop monitoring, suggesting potential UAV flight-time savings through image acquisition at higher altitudes.

0.25

The World as Viewed by 136 Countries: An Open, Curated, Annotated, and Multilingual Dataset of 118 Million State Media Articles

cs.CL cs.CY ·2026-10-18 ·Bryan Li, Samar Haider, Grace Kim, Fiona Luo, Jinqi Luo +4 more
datasetcode/data
Abstract

State media outlets often project their governments' narratives to target both domestic and foreign audiences. Because this content is plentiful and free to access, it appears widely in the web corpora that large language models (LLMs) are trained on and that retrieval systems draw from. However, the prevalence of state media in these corpora and its potential influence on LLM outputs remain largely unexplored. In this paper, we introduce WorldStateMedia, an open dataset of 118 million state media articles from 136 countries in 114 languages. We extract these articles from existing pretraining corpora and annotate each outlet with level of state control from State Media Monitor. Using WorldStateMedia to audit these corpora, we find that, in English web crawls, the share of state media pages from editorially controlled (vs. independent) outlets rose from 39% in 2013 to 93% in 2025. In several low-resource languages, state media make up a substantial share of all web data, such as 17.8% for Uyghur and 7.8% for Belarusian. To characterize the differing narratives of news outlets, we compare Russian and Ukrainian state media in English and in their native languages, revealing asymmetric topical focus and opposing framings of contested issues such as Crimea. Furthermore, in a retrieval-augmented generation (RAG) experiment on territorial disputes, retrieving a country's state media shifts model answers toward that country's claims, most strongly for English-language (foreign-targeted) media. We release WorldStateMedia to support multilingual, cross-national insights on state media at https://huggingface.co/collections/manestay/worldstatemedia

0.24

MoCKEFA: Model-Conditioned Keyframe Extraction and Fair Attribution for Video Model Explanation

cs.CV ·2026-10-18 ·Lahiru Wijayasingha
Abstract

We show that a video can be represented with just a few keyframes without sacrificing a given model's belief about the video. Our method is the first of its kind that extracts these keyframes conditioned on the model's prediction, which is massively helpful for building explainable video models. We then fairly distribute the importance of the video to the model among the keyframes with game theoretically fair Shapley values. Through qualitative and quantitative experiments, we show that keyframe attribution is superior to the traditional frame attribution. Further, we conduct preliminary exploration on highlighting the important changes among the frames by looking at the differences between adjacent keyframes.

0.24

FineSplat: Fine Patches, Local Groups, Detailed Splats

cs.CV ·2026-10-18 ·Hyebin Kim, Hwasik Jeong, Joonhyuk Park, Gyeongjin Kang, Jungwoo Kim +2 more
Abstract

Feed-forward 3D Gaussian Splatting (3DGS) promises instant scene reconstruction with high-quality novel-view synthesis. Yet, fine-grained geometry and appearance remain difficult to recover, motivating us to revisit how feed-forward models represent, aggregate, and refine detailed visual information. We introduce FineSplat, a feed-forward 3DGS framework built around fine-grained image patches. Since global multi-view attention becomes prohibitively expensive at such high token resolutions, we develop a dual-grouping strategy that combines two complementary forms of locality: view-wise grouping enables efficient interaction among neighboring views, while 3D spatial grouping reconnects geometrically related tokens that may never have interacted under view-wise grouping by lifting them into the 3D scene space. We further leverage reconstruction error as direct feedback, providing an explicit rendering-based signal to recover fine-grained structures and appearance, thereby overcoming the limitations of one-pass feed-forward prediction. Finally, a compact Gaussian generation scheme preserves the benefits of fine-grained features without unnecessarily increasing the number of primitives. Together, these designs enable FineSplat to recover richer geometry and appearance while remaining efficient and compact, achieving state-of-the-art results by a substantial margin, under the same or smaller Gaussian budgets than existing methods.

0.24

Where and Why Search-Augmented LLMs Fail When Reasoning over Rapidly Evolving Knowledge

cs.CL cs.AIcs.IRcs.LG ·2026-10-18 ·Tianyi Hu, Bhavyajeet Singh, Niket Tandon, Akhil Arora
benchmark
Abstract

As the world changes, new information is published while older records remain online. This complicates reasoning for large language models (LLMs), which must retrieve evidence and resolve conflicts to determine which information remains valid. Existing evaluations treat information seeking as a black box, revealing little about where and why LLMs fail. To address this gap, we introduce TREK, a suite of three components: TREKTRACE is a plug-and-play framework that attributes failures to specific stages of retrieval and reasoning. We apply it to TREKBENCH-WILD, a meta-benchmark of real-world questions that require complex reasoning over rapidly evolving knowledge. Surprisingly, the main bottleneck is post-retrieval reasoning rather than retrieval: frontier models are often misled by conflicting evidence even after retrieving all the information needed to answer. To study this bottleneck, we build TREKBENCH-CTRL, a controlled benchmark with distractors spanning temporality, relevance, and trustworthiness, and show that benchmarks dominated by outdated distractors substantially underestimate task difficulty. Together, TREK provides a testbed for diagnosing and improving reasoning over rapidly evolving knowledge.

0.24

PolarVis: A Dataset and Diffusion Framework for RGB-to-Polarization Perception in Adverse Driving Scenes

cs.CV ·2026-10-18 ·Ruilin Huang, Jue Wang, Zhen Tian, Xia Hua, F. Richard Yu
Abstract

Reflections, glare, transparent structures, and material-dependent appearance variations can introduce ambiguity into RGB-based visual perception in vehicle-relevant environments. Polarization provides complementary reflection- and material-sensitive information, but its acquisition requires dedicated sensing hardware. This paper presents RGB-P Wild, a real-world dataset containing 10,000 paired RGB and four-angle linear-polarization image sets, together with PolarVis, a diffusion-based framework for recovering polarization representations from a single RGB image. PolarVis combines hierarchical RGB feature extraction, diffusion-based feature refinement, four angle-specific decoders, and multi-scale attention recalibration to reconstruct polarization intensity images at $0^\circ$, $45^\circ$, $90^\circ$, and $135^\circ$. On RGB-P Wild, PolarVis achieves a normalized-domain MSE of 0.0465, a PSNR of 27.83 dB, and an SSIM of 0.8597. Compared with the released PolarAnything model under the adopted evaluation protocol, PolarVis obtains an AoLP error of $25.66^\circ$ and a DoLP MAE of 0.0404, compared with $46.03^\circ$ and 0.1084, respectively. Cross-dataset evaluation on vehicle-mounted imagery further examines transfer beyond the training domain. In downstream evaluation, a frozen PCDNet achieves AP50/AP50:95 of 86.3%/59.1% on 990 RGBP-Car test images using PolarVis-generated polarization, compared with 87.0%/59.8% using measured polarization. These results indicate that RGB-derived polarization retains useful polarimetric information for the evaluated perception tasks, while remaining distinct from calibrated physical polarization measurements.

0.24

Rubrics Carry the Standard: Shapley Attribution for Checklist-Based LLM Judges

cs.CL ·2026-10-18 ·Saeid Asgari, Rui Ying, Sara Malvar, Rakshanda Agarwal, Bruce Sun +3 more
Abstract

Checklist rubrics applied by large language model (LLM) judges have become a standard instrument for evaluating open-ended outputs and for rewarding them during training. Two assumptions surround this practice: that leniency or strictness is a property of the judge model, and that a rubric can be improved by asking a model to review it. We test both in a controlled grading task with planted flaws and on two human-labeled datasets, SummEdits and RAGTruth. Holding the judge and the outputs fixed, rubric content moves the pass rate of correct outputs from 2% to 100%, whereas changing the judge model or the rendering of the same checks moves it by a few points. Model self-review then drifts the rubric toward the reviser's own standard: it adds requirements the task never stated and deletes checks that human labels had taught, with three reviser models and on all three datasets. To measure and control this, we show that under all-must-pass aggregation the Shapley value of each check has an exact closed form, which yields per-check catch credit, false-fail blame and unique catches in linear time. Attribution shows that self-review deletes the checks that do unique work and keeps the checks that cause false fails. A change-control gate that combines these quantities with a provenance freeze on label-derived checks holds nearly every drifting revision chain at its starting quality, with identical decisions under three judge models.

0.24

Edge Deployable Lightweight Framework for Real-Time Driver Safety System with Integrated Lane Departure and Traffic Sign Recognition

cs.CV ·2026-10-18 ·Hassaan Ali Qureshi, Danish Hussain, Ali Hassan, Aon Safdar, Anas Bin Aqeel +3 more
published
Abstract

Lane departure and traffic-sign negligence are leading contributors to road accidents, particularly in regions with non-standardized signage and limited annotated data. Existing Advanced Driver Assistance Systems (ADAS) solutions are often developed as disjointed submodules, are rarely validated for real-time operation, and struggle to balance accuracy with low-latency inference for embedded deployment. We address these gaps with a unified, edge-deployable ADAS framework that simultaneously performs lane-departure warning (LDW) and traffic-sign detection and recognition (TSD/TSR), and validate the proposed system on automotive-class devices. Our framework combines a lightweight, lane-segmentation-driven LDW, which includes a drivable-area fallback for degraded markings, with a two-stage TSD/TSR pipeline consisting of efficient sign localization followed by fine-grained classification, designed to preserve accuracy at embedded latency. To support region-specific deployment under data scarcity, we curate a 35-class local traffic-sign dataset and introduce a generalized augmentation and balancing profile, expanding the corpus from 359 images to 8,750 samples. On an NVIDIA Jetson Orin Nano, the integrated system sustains 15 FPS end to end (32 FPS for LDW and 18 FPS for TSD/TSR). The TSD/TSR branch achieves 97.4 mAP for detection and 95.6% validation accuracy (93.1% on a local real-world set) while maintaining an approximately 13.9 M parameter footprint. Relative to recent methods that target a single task or report results only on desktop-class GPUs, our approach delivers both LDW and TSD/TSR with real-time performance on edge hardware, providing a practical pathway to low-cost, field-ready ADAS in resource-constrained settings.

0.24

ReTrace: Visual Backtracking for Efficient Multimodal Large Language Models

cs.CV cs.AI ·2026-10-18 ·Guangzong Si, Dong Wang, Yifan Yu, Miao Liu, Panwang Pan +1 more
published
Abstract · Accepted to Findings of EMNLP 2026, forthcoming

Multimodal Large Language Models (MLLMs) encode images into dense visual token sequences, creating substantial attention and KV-cache overhead. Existing visual token compression methods reduce this cost by pruning low-saliency tokens in shallow layers, but this practice couples temporary computational deactivation with irreversible information deletion. We show that this coupling breaks an intrinsic inference pattern in MLLMs. Through layer-wise attention analysis, we identify \textbf{Visual Backtracking}: visual pivots remain spatially stable, yet in middle layers they actively retrieve context from tokens that appeared unimportant earlier. Thus, early low-saliency tokens can be inactive for computation while still being useful for later reasoning. Motivated by this finding, we propose \textbf{ReTrace}, a backtracking-aware compression framework that introduces a dormant token state. ReTrace caches non-pivot tokens outside expensive layer-wise computation, restores them during the backtracking window, and prunes only after their information has been absorbed. Experiments on Qwen2.5-VL and InternVL3 demonstrate nearly 90\% reductions in FLOPs and KV cache, a 2.83$\times$ prefilling speedup, and strong performance retention across complex vision-language benchmarks.

0.24

Learning Lane-Structured Conditioning Fields for Online HD Mapping

cs.CV ·2026-10-18 ·Andrea Bertogalli, Anthony Dell'Eva, Marcello Ceresini, Federico Pirazzoli, Lorenzo Cipelli +3 more
code/data
Abstract

High-definition maps provide critical structural information for downstream autonomous driving tasks. Recent methods increasingly rely on external priors to improve prediction accuracy and robustness. Yet, existing prior representations expose a trade-off: dense rasters are easy to integrate but ignore certain map geometric properties, while sparse vectors preserve geometric structure but require dedicated encoders and are difficult to plug into existing architectures. Moreover, external priors can be inaccurate, costly to acquire, or difficult to maintain, motivating a structural representation that can provide geometric guidance while remaining learnable and broadly applicable. We address this limitation with the Lane Field, a dense BEV-aligned geometric representation that encodes local lane structure at each spatial location and can be either predicted from onboard observations or constructed from external map priors. Building on this representation, we introduce a plug-and-play framework that enhances online HD mapping through two complementary mechanisms: BEV feature modulation, which injects lane-structured cues into latent BEV features, and deformable attention guidance, which steers query sampling toward the underlying lane geometry. We primarily evaluate our method on OpenLane-V2, showing that conditioning BEV features and guiding deformable attention with local lane geometry enables our framework to consistently improve existing methods and achieve state-of-the-art performance on lane segment perception. Project page: https://andberto.github.io/LaneField/

0.23

Mixture-of-Memories: Low-Overhead Language Modeling with Modular Knowledge Injection

cs.CL ·2026-10-18 ·Yifan Lu, Rubin Wei, Ziyun Zou, Junhong Liang, Haolin Yang +6 more
Abstract · 18 pages, 8 figures

Large language models store substantial factual knowledge in their parameters, yet this knowledge is difficult to control, update, and reliably elicit at inference time. Existing memory-augmented approaches typically rely on a single auxiliary mechanism, such as retrieval, parameter updates, or neural memory, and therefore apply the same form of correction to tokens with very different knowledge demands. We propose Mixture-of-Memories (MoM), a memory-augmented language modeling framework that separates routed neural memory from writable factual memory. MoM combines a frozen base LM with two neural memory experts: a latent memory that provides low-latency hidden-state corrections and a lexical memory that provides a token-space knowledge pathway. A lightweight Budget Router sets per-token mixture weights from uncertainty and disagreement features, replacing the fixed interpolation of prior memory modules. For post-hoc updates, an external explicit memory stays outside the router, because a router trained before facts are written learns to down-weight it; new facts are appended without retraining any component and used only when an answer-level gate accepts the retrieved fact. Experiments across language modeling, factual QA, update analysis, and efficiency show that MoM improves both predictive accuracy and updateability while retaining low inference overhead.

0.23

Sequential Transfer Learning for Plant Disease Recognition Across Perennial Crop Domains Using MobileNetV2-CBAM

cs.CV ·2026-10-18 ·Ropkathaa Choudhury
benchmark
Abstract · 17 pages, 5 figures, 5 tables

Purpose: This study investigates whether a learned disease-recognition representation can be progressively reused across distinct perennial crop domains and how its predictive behaviour changes as target-domain heterogeneity increases. Design/methodology/approach: A lightweight MobileNetV2 framework enhanced with the Convolutional Block Attention Module (CBAM) was evaluated through sequential transfer from Walnut to Cashew, Coffee and Mango, followed by validation on the heterogeneous ABCGMP six-crop benchmark. Each stage used the preceding model representation and was evaluated on an independent test set using complementary classification measures. Findings: Test accuracy increased from 75.76% for the Walnut baseline to 84.85% after CBAM integration, followed by 90.67% on Cashew, 100.00% on Coffee and 99.50% on Mango. The final ABCGMP evaluation achieved 88.44% accuracy, 81.49% Balanced Accuracy and 0.9949 ROC-AUC. The results show that substantial predictive capability was retained through successive crop transfers, while broader multi-crop heterogeneity introduced a measurable performance constraint. Originality: The study provides an empirical characterisation of sequential knowledge reuse across multiple perennial crop domains using a common lightweight attention-enhanced representation, extending evaluation beyond isolated source-to-target transfer.

0.23

SPARC: Spoken Profile Authorship Recognition Corpus

cs.CL ·2026-10-18 ·Efthymios Tsaprazlis, Bowen Yi, Tiantian Feng, Junghwan Kim, David Jurgens +1 more
benchmark
Abstract

Speech anonymization has focused on the acoustic signal, leaving open how much speaker information the transcribed words themselves retain. In this work, we study pairwise speaker and demographic-attribute verification from transcribed speech. We introduce SPARC (Spoken Profile Authorship Recognition Corpus), a benchmark built from two podcast corpora comprising over 12,500 speakers, 1.5M quality-filtered utterances, and speaker-level inferred or source-provided labels for sex, age, and accent, and evaluate five attribution models and three LLMs on identity- and attribute-verification trials with episode- and metadata-based topic proxies. LLM adversaries distinguish same-speaker pairs from text alone. Attribute-label agreement is harder to verify, but LLMs perform sex- and accent-label verification above chance. Finally, releasing only short probe texts (50 words) degrades the strongest adversaries the most, offering a partial reduction in observed linkage.

0.23

ATLAS: Attention-Guided Test-Time Latent Selection for Compositional Visual Autoregressive Generation

cs.CV ·2026-10-18 ·Hossein Shahabadi, Raya Namazi, Mahdieh Soleymani Baghshah
Abstract · 18 pages

Visual autoregressive models produce convincing images but can misassign attributes and spatial relations. Searching multiple generation paths can improve composition, yet completing every candidate is costly. We introduce ATLAS, a training-free framework that uses the frozen generator's internal attention to guide search before final images are available. Within-prompt diagnostics identify attention-derived scores that predict final compositional preferences on Infinity, with category-dependent transfer to Switti. ATLAS uses these on-the-fly signals to progressively concentrate computation on promising trajectories. It maintains multiple coarse alternatives, prunes candidates without decoding at intermediate scales, and reserves image-space evaluation for a final choice among partial trajectories. ATLAS scales effectively on both backbones, achieving higher mean scores on the evaluated T2I-CompBench and GenEval categories as more computation is devoted to search. At the largest tested budget, the targeted T2I-CompBench mean improves by over 10% relative to each backbone. Separate profiles show over 50% less generation time than completing and evaluating all candidates at equal initial counts, alongside lower peak allocated GPU memory. Relative time savings are larger at the higher tested candidate count. Together, these results support scalable compositional search: additional sampling expands exploration, while attention-guided selection limits continuation cost without parameter updates or gradient-based optimization.

0.23

Do LLMs Reproduce the Age Gap in Loneliness? Evaluating Survey Simulation against European Survey Data

cs.CL ·2026-10-18 ·Yuliya Li, Achim Rettinger, Dominik Beyer
unusual
Abstract

Large language models (LLMs) are often used to simulate survey respondents, but matching survey averages does not establish accurate demographic differences. We test whether LLM outputs reproduce within-country age gaps in loneliness using the 2022 EU Loneliness Survey, comparing adults aged 18-34 and 65 or older in France, Italy, Poland, and Sweden. We evaluate simulation with and without age information and direct distribution estimation against a probability-based sample, with a non-probability sensitivity reference. Both human samples show higher mean loneliness among younger adults. Under age-specified simulation with standard response-option order, Mistral Small 4 and GPT-5.6 Terra reverse this ordering under all three prompt framings in three and two countries, respectively, mainly by overestimating older-adult means. DeepSeek V4 Flash preserves the ordering everywhere but underestimates both group means, with errors largely cancelling in France and Italy. Against the probability-based reference, direct estimation has lower absolute age-gap error than simulation for Mistral and Terra in every country. For DeepSeek, simulation has lower gap error but direct estimation has lower group-mean error; these comparisons depend on low-validity older-adult simulation cells. Evaluating LLM outputs as proxies for age groups within countries therefore requires checking both group means and their contrast against human survey data.

0.22

DragonMoCap: Monocular Video-based Motion Capture for Diverse Dragon Topologies

cs.CV ·2026-10-18 ·Zhongfei Qing, Xiaoxiang Zhao, Ling-Hao Chen, Zhongang Cai, Wenjia Wang +5 more
Abstract · Siggraph Asia 2026

This work explores monocular motion capture for dragon-type characters, which currently relies on manual keyframing due to the lack of high-quality datasets and parametric models. The primary obstacle lies in the extreme topological diversity across dragon rigs, where varying joint counts and naming conventions render traditional fixed-template regressors inapplicable. To address this, we introduce DragonMoCap, a novel framework that treats the target rig as a dynamic input rather than a fixed prior. By deliberately scoping the framework to the most prevalent production topologies, we ensure high-fidelity results across diverse morphologies. The core of our approach is the design of language-grounded joint query tokens, which establish a dual-property representation: ensuring intra-rig discriminability while maintaining cross-rig similarity. We further propose a non-parametric query-token retrieval mechanism to enable zero-shot adaptation to unseen rigs without retraining. Extensive evaluations demonstrate that DragonMoCap consistently outperforms HMR-style baselines across diverse dragon topologies. The practical utility of our framework is further evidenced by its integration into professional animation workflows, highlighting its potential for real-time industrial applications.

0.22

Evaluating Cloud and Self-Hosted OCR Systems for Arabic, English, and French Documents: Bounding Box Detection, Text Recognition, and Mathematical Expression Recognition

cs.CV ·2026-10-18 ·Jumana Alsubhi, Ahmed Alhusayni, Abdulrahman Gharawi, Israa Hamdine, Khadija Alansari +3 more
Abstract

Choosing an OCR system for a multilingual archive requires balancing accuracy against deployment constraints. This study compares cloud and self-hosted systems on three tasks: bounding box detection on 48 Arabic, English, and French pages, text recognition on 315 pages across nine document types, and mathematical expression recognition on 83 English and Arabic samples. Detection was assessed by page-level visual grades, text by character error rate (CER) against a Gemini-generated reference transcription, and mathematics by scores against known answers. Main estimates are accompanied by 95% intervals and paired comparisons. Google Document AI and Surya produced Good line boxes on 47 of 48 (98%) and 46 of 48 (96%) pages. Datalab, Azure Document Intelligence, and Document AI had mean CERs of 0.145-0.152, with no difference detected in unadjusted paired comparisons. Surya had the lowest mean CER among self-hosted systems (0.199). No mathematical test had a statistically clear leader. Strict checks showed that near-perfect page scores could conceal a replaced Arabic variable or Eastern Arabic digits rewritten in Western form. Reliable evaluation therefore requires uncertainty estimates, structural checks, and notation-specific tests in addition to aggregate recognition scores.

0.22

A New Hope for PRNU on Modern Smartphones? Evidence That Device-Specific Fingerprints Survive Computational Photography

cs.CV eess.IV ·2026-10-18 ·Benjamin Loison, Quentin Bammey, Thibaud Ehret, Rafael Grompone von Gioi, Pablo Musé +1 more
unusual
Abstract

Photo Response Non-Uniformity (PRNU) is widely regarded as the most reliable marker for source camera identification. Recent studies, however, report high correlations between images and fingerprints from different modern smartphones, which calls into question the discriminative power of PRNU on these devices. Using 30 recent smartphones under controlled acquisition conditions (e.g., no digital zoom), we investigate whether the fingerprint itself is at fault or the way it is compared. We find that a component shared by all fingerprints of the same model inflates the standard signed peak-to-correlation energy (sPCE). To counter this, we propose $Δ$sPCE: the sPCE between a query image and a candidate camera's fingerprint, minus the mean sPCE between that image and the fingerprints of the remaining candidates, in the spirit of T-norm score normalization. $Δ$sPCE improves attribution accuracy, suggesting that the device-specific PRNU signal is still present. Although the limited dataset size and the dependence of $Δ$sPCE on the candidate camera set currently restrict operational applicability, these results suggest that PRNU-based attribution of modern smartphones remains feasible when fingerprints are compared appropriately.

0.22

Assertion and Belief: Moore's Paradox in Language Models

cs.CL ·2026-10-18 ·Anshun Asher Zheng, Weili Cao, Fangcong Yin, David Beaver, Junyi Jessy Li
unusual
Abstract

Moore's paradox \citep{moore1942} concerns sentences such as ``It is raining, but I don't believe that it is,'' which may well be true yet cannot be sensibly asserted. The absurdity lies not in the content but in the act, since to assert is to vouch for one's own belief. This makes the paradox a test of whether language models take the speech act into account when computing meaning. We present the first test of Moore's paradox in LLMs, evaluating 12 models on a curated set of Moorean sentences in varied contexts. We find that models largely recognize the infelicity of Moorean sentences, and that the judgement is not tied to a memorized form and tracks the speech act. Using the Rational Speech Act framework (RSA; \citealt{goodmanfrank2016}), we then test whether these judgements arise from a listener reasoning about a speaker. We find the models' own listeners do not reason about their own speakers, and their reasoning traces suggest that they instead recall knowledge about the paradox and apply rules therein. The models thus appear to have learned the paradox as a fact about sentences, not as a requirement on speakers.

0.22

TOMORI: A Tool-Orchestrated Mixing Operator for REAPER Interaction

cs.SD eess.AS ·2026-10-18 ·Yining Wang, Tianjiao Li
Abstract · The authors are withdrawing this preprint at the request of their company

Code-based agents can operate creative software, but mixing requests often describe an intended sound without specifying the edits needed to achieve it. We present TOMORI, a code-based mixing-agent harness for native digital audio workstation (DAW) control. Custom plugins provide REAPER project inspection, optional audio measurement, and execution of model-generated Lua code. All observations return as text; delivery includes an editable project and rendered audio. TOMORI-BENCH pairs 110 human-reviewed reference edits with three instruction forms and evaluates native DAW execution end to end through operation structure, parameter precision, and rendered-audio outcomes. Across 990 cases from three primary models, explicit instructions yield high reference agreement, while abstract requests reduce project agreement and audio improvement. Paired ablations show that the full workflow improves explicit execution over a single submission. Prompted audio measurement succeeds in all 28 selected abstract cases but yields no clear overall improvement on the reference-based metrics and increases tool calls. These results expose the gap between executable DAW control and precise realization of perceptual mixing goals.

0.22

Walk in Envelopes: Efficient Branching Walk-on-Spheres via Certified Solution Bounds

cs.GR ·2026-10-18 ·Yunfeng Liao, Zhenhao Qiu, Xinyang Chen, Xiucheng Li
Abstract

Branching Walk-on-Spheres (WoS) solves semilinear PDEs without meshes, but its computational cost grows sharply near critical branching rates. We introduce Walk in Envelopes (WiE), which uses certified solution bounds to transform branching diffusion rather than prune existing trees. The transformation suppresses reproduction, introduces controlled termination, and preserves unbiasedness with an envelope-dependent variance bound. We realize the transformed process through exact rejection sampling using standard screened WoS kernels, supported by certified spline envelopes and adaptive sampling, and further prove that suitable harmonic envelopes yield bounded expected reproduction counts even as conventional BBM diverges.

0.22

LC-TSRNet: A Land-Cover-Guided Two-Stage Network for Annual MODIS Surface Reflectance Time Series Reconstruction

cs.CV ·2026-10-18 ·Wenjuan Zhang, Xuesong Ma, Zhen Li, Meixiang Wang, Caimeng Yi +3 more
Abstract

MODIS surface reflectance (SR) products often suffer from data gaps due to cloud contamination and sensor malfunctions, limiting their applications in long-term and large-scale studies. Most existing methods rely on cloud-free observations from optical time series or introduce SAR time series as auxiliary information. However, due to the coarse spatial resolution of MODIS, cloud-free reference images are difficult to obtain, and SAR imagery is not well matched to the spatial resolution of MODIS, making it difficult to provide effective prior constraints. To address these issues, this study proposes LC-TSRNet for seamless reconstruction of annual MODIS SR products. In the first stage, an annual land cover map is introduced as prior knowledge, and CSPM is designed by leveraging the spectral similarity within identical land cover types, thereby providing prior spectral constraints for missing regions. In the second stage, to address large-area, persistent, and overlapping gaps in MODIS imagery, SMRM is designed to adaptively exploit and integrate multi-scale spatiotemporal contextual information. Meanwhile, to address the limited collaborative utilization of spatio-temporal-spectral information, 3DAM is designed to impose cross-dimensional joint constraints. Considering the lack of publicly available MODIS sample datasets, this study constructs the AIRMOD-TSR dataset. Extensive experimental results show that LC-TSRNet outperforms the comparison methods and achieves better reconstruction accuracy and spectral fidelity. Under the spatially adjacent scenario, LC-TSRNet improves PSNR by 0.56 dB and reduces SAM by 0.13°. The vegetation monitoring application further verifies the practical value of the reconstructed SR time series.

0.21

Local Evidence, Global Failure: Diagnosing Speech-Vision Temporal Composition in Audio-Visual Language Models

cs.CV cs.AIcs.MM ·2026-10-18 ·Jaehong Myung, Seunghun Oh, Unsang Park
Abstract

Audio-visual large language models (AV-LLMs) can fail at full-video speech-vision decisions even when useful visual evidence is locally recoverable. We introduce Temporal Evidence Decomposition (TED), a training-free diagnostic framework that separates local visual verification from speech-relative aggregation. Automatic speech recognition supplies a transcript and utterance interval. TED queries audio-free windows with the transcript for Temporal Alignment and Forecasting, or answer candidates for Cross-Modal Temporal Binding, then aggregates evidence relative to that interval. Across four frozen AV-LLMs on SVHalluc, Temporal Forecasting accuracy rises from 29.3-33.7% under full-video questioning to 79.3-82.2% with TED. Query replacement reduces recovery; the observed anchor remains beneficial when temporal-region availability is held fixed. Physical-isolation and timing-shortcut controls test alternative explanations. The findings distinguish locally recoverable evidence from successful full-video decisions, motivating explicit separation of narration content, utterance timing, and time-indexed visual evidence.

0.21

LAMB: Latent Multi-Hypothesis Beliefs for Long-Horizon Large Language Agents

cs.CL ·2026-10-18 ·Peien Ding, Hui Li, Yuhan Liu, Jinsong Su
Abstract

Long-horizon language agents require compact memories that retain interaction evidence while preserving competing hypotheses about unobserved states. Although effective, many belief-state methods encode memory in discrete natural language, where repeated compression can omit evidence or prematurely discard plausible alternatives. We introduce LAMB, a recurrent framework for latent multi-hypothesis beliefs. A Common--Delta factorization combines a shared Common state for interaction evidence with weighted Delta components for competing hypotheses. Relative evidence from action-outcome predictions guides recurrent hypothesis updates without generating textual memory. A context-conditioned readout exposes hypotheses separately to the policy through a fixed-token belief prefix in the default configuration. The belief modules and policy are jointly trained using action imitation and auxiliary state, outcome, and observation-discrimination objectives. Across ALFWorld, BabyAI, ScienceWorld, and Crafter, LAMB matches or outperforms PABU, improving ALFWorld SR@30 and ScienceWorld success by 6.0 and 22.0 percentage points, respectively, and Crafter achievement progress by 26.37 points. On ALFWorld, it also reduces the redundant-action rate by 17.70 percentage points. These results support latent multi-hypothesis beliefs as compact memory for effective long-horizon decision making

0.21

MemBubble: Benchmarking How Agent Memory Shapes LLM Response Diversity

cs.CL ·2026-10-18 ·Jiajun Tan, Fei Sun, Xinyu Ma, Shixuan Zhang, Du Su +4 more
benchmark
Abstract

Memory serves as the core component of LLM-based agents, enabling personalized responses via extracting and using user preferences from past interactions. However, its use may reduce response diversity, thereby narrowing the range of information available to users. Existing evaluations rely largely on synthetic interactions and cover few memory systems, leaving the effects of memory on response diversity insufficiently evaluated in real-world scenarios. We introduce MemBubble, a benchmark of 1000 instances built on preferences extracted from real user-assistant conversations. To measure the extent to which memory concentrates response distributions, we construct test queries in scenarios distinct from user preferences while preserving topical relevance through shared preference categories. Across multiple LLMs and memory mechanisms, entropy decreases by 34.7% on average compared with directly providing the full interaction history, highlighting the important role of memory processing in diversity loss. We also explore simple user-side interventions and changes in information availability but find that these do not effectively close the diversity gap. Our findings highlight the need to consider preserving diversity and user access to information alongside improvements in personalization when designing and using agent memory systems.

0.21

UniState: Unifying Memory and Intervention for Controllable World Simulation

cs.CV ·2026-10-18 ·Jaeseok Jeong, Kyungmook Choi, Sohyun Chung, Youngsik Yun, Youngjung Uh
code/data
Abstract · Project page: https://curryjung.github.io/unistate/

World models offer increasingly rich environments for exploration, yet intervening in a world while observing the consequences from freely chosen viewpoints remains challenging. We tackle these challenges by introducing UniState, a framework for controllable world simulation built on a unified world-state representation. This representation unifies spatial memory and user interventions. Users specify object motion directly within spatial memory, while independently choosing the camera trajectory from which the resulting world state is observed. This unification makes the relationship between the scene and the requested changes explicit, removing ambiguity between object motion and viewpoint change from the control specification. Crucially, specifying an intervention does not require prescribing every resulting movement: the video model can synthesize additional motion within controlled objects and plausible interactions with their surroundings. We develop training and sampling strategies that help generated videos follow the specified object motion and camera trajectory while preserving this capacity to generate responses beyond the prescribed transformations. Experiments demonstrate more accurate joint object and camera control than the existing methods, alongside qualitative evidence that intervention-induced scene changes persist across rollouts.

0.21

Image Restoration with Impulse Noise Based on Minimax Concave Penalty: Grayscale and Color Images

eess.IV ·2026-10-18 ·Reza Parvaz
Abstract

In today's world-where social networks play a significant role-images serve as one of the most important tools for transferring concepts. Image degradation caused by noise or blurring hinders the effective transmission of the information we expect to derive from an image. Consequently, developing image restoration methods capable of revealing crucial information is of great importance. In this work, a model for impulse noise image restoring is proposed by using a Minimax Concave Penalty. The proposed model is applicable to both color and grayscale images. Then, a numerical algorithm based on the alternating direction method of multipliers is introduced, and the convergence of the method is analyzed both theoretically and numerically. In the final step, the proposed method is evaluated through simulations on several examples, and the results demonstrate its effectiveness.

0.21

CoRe-WM: Conditioned Response-State World Modeling for Longitudinal Treatment-Response Prediction in Breast DCE-MRI

cs.CV ·2026-10-18 ·Lin Zhao, Bowen Chen, Qiaohui Gao, Shaowen Wan, Xiang Li +2 more
Abstract

Pathologic complete response (pCR) after neoadjuvant chemotherapy is a key prognostic indicator in aggressive breast cancer, yet existing prediction approaches either rely on pretreatment imaging alone, missing the patient's evolving response, or optimize multi-visit classifiers under direct outcome supervision, coupling representation learning with the clinical endpoint. We formulate longitudinal pCR prediction as a world-modeling problem and propose Conditioned Response-state World Modeling (CoRe-WM), which learns a latent state-transition model that predicts the next imaging state from the observed sequence. The transition function is conditioned on patient clinical profile and assigned therapy through additive shift and feature-wise modulation, enabling patient- and treatment-specific dynamics. A Factorized Response State pathway further decouples treatment-response dynamics from imaging appearance by encoding the structural prior that response manifests primarily in lesion morphology. After pretraining, all parameters are frozen and a linear readout predicts pCR at each clinical decision point. We evaluate CoRe-WM on I-SPY2 under five-fold cross-validation and assess cross-cohort transfer on I-SPY1. CoRe-WM achieves the highest discrimination among all evaluated methods at every prospective decision point, and demonstrates robust cross-cohort transfer. Ablation studies identify patient-profile conditioning and the factorized response pathway as the principal contributors. These results establish world-model-based representation learning as a promising framework for outcome prediction in oncology and beyond.

0.20

ReDT: Resolution-Decoupled Visual Geometry Transformer

cs.CV ·2026-10-18 ·Abteen Arab, Guile Wu, Dongfeng Bai
Abstract · Technical report

We introduce ReDT, a feed-forward visual geometry transformer designed to scale multi-view reconstruction to high-definition (HD) inputs. Existing visual geometry transformers are typically trained at fixed, standard definition (SD) resolutions and exhibit degraded reconstruction quality and poor computational scaling to high input resolutions. While recent efficient architectures reduce the complexity of inter-frame attention, expensive intra-frame and encoder attention remain largely unaddressed, making HD reconstruction impractical with existing architectures. ReDT addresses both limitations with a resolution-aware architecture designed to scale to high-resolution inputs. The key idea of ReDT is to first map each view to a fixed-size set of basis tokens, compute correspondences in this compact constant-size basis space, and then project the outputs back to the original resolution for dense prediction. By decoupling the internal representation from input resolution during attention, ReDT substantially reduces the computational cost of high-resolution 3D reconstruction and improves model performance at all resolutions. Extensive experiments on multiple datasets spanning SD to FHD (2K) show that ReDT outperforms state-of-the-art visual geometry transformers in both reconstruction quality and computational efficiency. For instance, on the ETH3D (2K) dataset, ReDT achieves a $16\times$ speedup, a $34\%$ reduction in peak GPU memory usage, and a $33\%$ reduction in chamfer distance over VGGT-$Ω$. \href{https://redt.github.io}{Project Website.}

0.20

A Measurement Perspective on Benchmarking LLM-Generated Text Detectors using Item Response Theory

cs.CL ·2026-10-18 ·Yanbin Zhao, Zhipeng Chen, Guangsheng Bao, Guandong Zhang, Yan Li
benchmark
Abstract

Benchmarking large language model (LLM)-generated text detectors commonly relies on aggregate detector-level metrics such as accuracy, F1 score, and AUC. Although useful for summarizing overall performance, these metrics provide limited insight into how benchmark composition shapes detector comparisons or which texts distinguish detectors with different response profiles. We propose a measurement-based benchmarking framework for jointly characterizing benchmark heterogeneity and detector behavior using multidimensional item response theory (IRT). By treating benchmark texts as measurement items and detectors as respondents, the framework estimates text-level discrimination and difficulty-location parameters alongside multidimensional detector abilities. We apply the framework to a benchmark of 32K human-written and machine-generated texts across four domains, evaluated by 14 detectors. The estimated IRT item-parameter space reveals distinct but overlapping human-machine distributions, with substantial variation across domains and text lengths. The human-machine label shows the strongest multivariate association with this space, while domain and text length further contribute to benchmark heterogeneity. Detector ability profiles reveal multidimensional and domain-dependent response patterns that are not fully reflected in aggregate performance rankings. As a practical application, clustering texts in the estimated IRT space yields a compact subset of 2,088 texts (6.5\% of the full 32K benchmark) that largely preserves the comparative detector-performance patterns observed on the full benchmark. Overall, the proposed framework provides a principled and interpretable methodology for benchmarking heterogeneous detection technologies while jointly characterizing the measurement properties of the texts used for their evaluation.

0.20

MPIG: Multi-angle Polarization Image Generation from a Single Polarization Observation

cs.CV ·2026-10-18 ·Jue Wang, Ruilin Huang, Liquan Shen, Zhen Tian, Xia Hua
Abstract

Polarization imaging provides informative cues for material analysis, scene understanding, and visual perception. However, acquiring complete polarization observations typically requires specialized hardware and multi-angle measurements, thereby increasing system complexity and limiting practical deployment. In this work, we address the problem of predicting the missing $45^\circ$, $90^\circ$, and $135^\circ$ polarization views from a single $0^\circ$ observation by leveraging a learned scene prior. To this end, we propose MPIG (Multi-angle Polarization Image Generation), a Swin Transformer-based framework that couples cross-angle feature interaction with residual decoding to predict angle-dependent intensity variations while preserving shared scene structure. Specifically, measured Stokes differences and polarization-strength-aware supervision guide the recovery of polarization modulation, orientation, and magnitude, while low-rank factorization further reduces the decoding cost. We also construct HighPol, a large-scale real-world polarization dataset comprising 5,474 groups of high-resolution four-angle polarization images collected from diverse indoor and outdoor scenes. Experimental results demonstrate improved image and polarization accuracy, achieving 29.729 dB PSNR, 0.9505 SSIM, and $21.514^\circ$ AoLP MAngE while requiring 85.86 M parameters and 1.126 T FLOPs. Furthermore, downstream validation on glass and semantic-segmentation tasks suggests that the generated polarization images can preserve task-relevant information, demonstrating the effectiveness of MPIG and the practical utility of predicted polarization observations for downstream vision applications.

0.20

Memories That Outlive Their Reasons: Dependency-Tethered Memory for Tool-Using LLM Agents

cs.CL ·2026-10-18 ·Yiqi Wang, Jinqian Ju, Jiaqi Zhang, Zhangkai Wu, Yiqun Duan +3 more
Abstract

Agents now act on the Web, publishing packages, models and datasets through registry APIs, and remembering past verifications spares them the work. But a verdict rests on premises. When a signing key is revoked, everything built on it stops passing, yet the advisory names only the key, and the old verdicts of its dependents may never share a retrieval window with it. This non-locality of invalidation turns memory into a liability: in long simulated chats, an agent reading a memory of atomic facts publishes 62% of the objects whose failure lies upstream, each backed by a record that was true when written. Yet derivation is local even when invalidation is not. Tether records, at write time, the exact versions each verdict derives from, and at read time withdraws any verdict whose support has failed anywhere in the store; the agent reads the same fields as before. On Fallout, a generator of versioned worlds with hypothesis-blind simulated histories shared by every memory, Tether publishes 0.3% of upstream-failing objects across three domains when relations are stated with their verdicts, with 37% fewer verification calls than checking everything. A-Mem and Graphiti re-verify most candidates (0.83--0.86 checks per candidate), which limits their savings, yet publish 9% of upstream-failing objects; Mem0 relies on memory more and publishes 42%. Flushing every verdict on any failure is as safe as Tether but costs 32% more checks. The gain over fact memory holds with three open-weight backbones, a whole-store window and real Go advisories. These gains depend on capturing dependencies: when relations arrive in sessions apart from their verdicts, Tether's downstream unsafe publication rate rises to 9.4--22.1%. It also degrades with weaker writers that miss dependencies or invalidation events.

0.20

Representations in Motion: Layer-wise Rotation and Trajectory Divergence in LLMs

cs.CL ·2026-10-18 ·Zhuonan Yang, Yik Siu Chan, Wannan Yang, Ellie Pavlick
Abstract

Recent work suggests that activation dynamics--how representations change during inference--might be more revealing about language model behavior than static readouts from individual layers. In this work, we investigate the layer-wise evolution of representations of factuality, using true and false statements when the model is prompted to either lie or to be honest. We report a geometric phenomenon in which the factuality directions extracted from the lying and honest conditions appear to rotate away from each other across layers. We find that such layer-wise rotation also occurs in a task that requires the model to take different visual perspectives in the 3D world, suggesting that the rotation phenomenon might be indicative of manipulating conceptual spaces in general, rather than unique to factuality. We then investigate the "real time" transformation of conceptual representations on a trial-to-trial basis. We show that it is possible to make strong predictions about the model's target task and behavior from trajectories alone. In addition, diverging points in single-trial trajectories correspond to the onset of rotation in aggregated-trial analysis. Finally, we show that both rotation and divergent trajectories occur in more naturalistic settings in which models are not overtly prompted to "lie" but rather report false statements in the course of completing a different task.

0.20

GeoART: Geometry-Aware Refinement via Test-Time Scaling for Camera-Conditioned Video Diffusion

cs.CV ·2026-10-18 ·Liudi Yang, Ziyuan Liu
Abstract · 20 pages, 11 figures

Camera-conditioned video diffusion models often suffer from inaccurate camera control and geometric inconsisten- cies because camera constraints and estimated geometry are only weakly enforced during denoising. Existing reme- dies often require costly geometry-aware fine-tuning or ad- ditional 3D supervision. We introduce GeoART, a training- free self-refinement framework that improves camera con- trollability through geometry-aware test-time scaling. At selected denoising steps, GeoART generates multiple can- didates within the Predict-and-Perturb cycle and eval- uates them using a geometry prior, without directly modify- ing the denoising trajectory. A confidence gate suppresses unreliable geometric evidence, while rank-calibrated prob- abilistic selection combines score magnitude and relative rank to choose the next refinement state. Experiments demonstrate consistent improvements in camera controlla- bility, multi-view consistency, and visual quality over exist- ing sampling and self-refinement methods, without training or modifying the pretrained video diffusion model

0.20

P2N: Reusing Deep Representations for Greater Effective Depth

cs.CL ·2026-10-18 ·Yiqin Hao, Boyi Zeng, Zitong Wang, He Li, Shixiang Song +5 more
Abstract · 23 pages, 10 figures

Mathematical problem solving and autonomous agents are increasingly important applications of language models, demanding stronger reasoning capabilities. Computational depth is a key resource for such reasoning, but adding model layers increases both training and inference costs, while test-time scaling incurs substantial additional inference computation. We introduce P2N (Previous to Next), which reuses the previous token's deep representations, already available during generation, to increase effective computational depth with negligible decoding overhead. Specifically, P2N adds the previous token's deep hidden state to the current token's shallow representation, allowing subsequent layers to build on previously completed deep computation. This recurrence extends effective computational depth across tokens, while introducing only a single vector addition per decoding step. We train the recurrence using parallel Jacobi updates restricted to a middle-layer core. Experiments with Qwen-style models show that, at 1.5B parameters, P2N improves GSM8K accuracy by over 10 percentage points under both zero-shot and few-shot prompting, with nearly unchanged decoding latency. The benefits also extend to language modeling and average performance on general downstream benchmarks. on general downstream benchmarks.

0.19

Concept Representations in Vector Space: Toward an Interdisciplinary Study of Meaning Variation

cs.CL cs.CY ·2026-10-18 ·Simon Josef Durstewitz
code/data
Abstract · 20 pages, 5 figures. Companion paper to arXiv:2610.15496. Code: https://github.com/simon-josef/conrep

An increasing number of disciplines recognize the explanatory potential of meaning-bearing mental structures through which individuals interpret the world and guide their behavior. Concept representations constitute a core class of these structures, emerging from individuals' experiences with concept referents and encoding and organizing their understanding of them. Recognizing a private level of meaning, where concepts are not objectively defined by rules in the world but constitute subjective structures in individuals' minds, raises a fundamental question: do different individuals actually mean the same thing when they invoke the same concept? As latent structures, concept representations are difficult to access and model empirically. The present study introduces a measurement framework that makes such representations empirically accessible through textual traces elicited under standardized conditions. Using transformer-based sentence embeddings, the framework maps standardized responses elicited for concepts into a high-dimensional shared vector space, opening concept representations to the systematic analysis of interindividual variation. It develops three analytical avenues: (1) comparing representational sharedness across concepts, (2) identifying variables that correlate with representational variation, and (3) examining how focal actors or subpopulations deviate from a reference population. Each avenue is illustrated through an empirical proof-of-concept drawing on a large-scale free-association dataset. Together, the demonstrations show how researchers from across disciplines can move from a research question and data source to analysis and interpretation. An accompanying GitHub repository provides a ready-to-use implementation of the framework for existing and custom datasets.

0.19

RARE: Reinforcement Learning with Adaptive Rubric Evolution for Open-Ended Generation

cs.CL cs.LG ·2026-10-18 ·Xueyun Tian, Bingbing Xu, Chuyue Li, Jiaming Liu, Chao Zhang +7 more
Abstract

Verifiable rewards have advanced reinforcement learning in mathematics and code. However, many valuable open-ended tasks share three properties: no automatic verifier of overall quality, multiple requirements to satisfy jointly, and multiple valid responses to each query. Training therefore needs feedback on requirement satisfaction across different valid responses. Rubric-based reinforcement learning provides this feedback by aggregating scores from explicit criteria. As learning progresses unevenly, some criteria become fully satisfied and lose discrimination while others remain informative. Dynamic methods revise shared rubrics or select evolving criteria to restore discrimination, but may also change useful criteria. To address this, we propose Reinforcement Learning with Adaptive Rubric Evolution (RARE). The core idea is to use current policy feedback to guide both which criteria change and what their replacements assess. Sample-level updates replace criteria fully satisfied across current outputs with checks of remaining quality differences, retaining all other criteria and scores. Within this rule, we compare rewarding desirable qualities with penalizing observed flaws. Video script generation exemplifies these properties and supports content production. Its demands for coherent, executable narratives under constraints make it a challenging testbed. Targeting flaws performs best across three video types with a mean gain of 3.82 points over matched static rubrics. Model and human preferences support these gains. Improvements also extend to writing and general generation benchmarks.

0.19

RAHF-Net: Range-Aware Heterogeneous Fusion for LiDAR-Assisted RGB Semantic Segmentation

cs.CV ·2026-10-18 ·R. D. Lin, Pengcheng Weng, Ge Wang, Han Ding, Fei Wang
Abstract

This paper studies LiDAR-assisted RGB semantic segmentation, where LiDAR point-cloud features improve dense pixel-wise predictions in the RGB image plane rather than 3D point labels or detections. Existing 2D fusion methods often project point clouds into shallow depth/range maps and apply spatially uniform fusion, discarding rich 3D structure and ignoring the range-dependent reliability of outdoor LiDAR. We propose RAHF-Net, a Range-Aware Heterogeneous Fusion Network. Its Cylindrical-to-Perspective (C2P) module projects pretrained cylindrical voxel features into the image plane, preserving structured 3D semantic priors. A Range-Aware Heterogeneous Fusion (RAHF) module then uses a soft depth-derived range mask to perform LiDAR-guided geometric enhancement in reliable near-field regions and visual-context enhancement in sparse far-field regions. On KITTI-360, RAHF-Net achieves 65.76 mIoU, outperforming representative RGB-only and RGB-LiDAR baselines, including CMX by 1.45 points. Experiments on DELIVER further demonstrate improved robustness under adverse weather, illumination degradation, and sensor corruption.

0.19

The Explainability Illusion: Must ViTs Exhibit Patch-Decoding Interpretability?

cs.CV cs.LG ·2026-10-18 ·Ronen Schaffer, Hila Manor, Tomer Michaeli
Abstract · Code can be found in https://ronensc.github.io/Explainability-Illusion/

Transformers are widely assumed to exhibit token-decoding explainability. In particular, in vision transformer (ViT) classifiers, decoding individual patch tokens has become a common diagnostic tool due to its tendency to yield human-interpretable outputs. In light of the increasing popularity of this approach, an important question is whether token-decoding explainability is an inherent property of transformers that should be expected to hold across architectures, scales, and training objectives, or merely an artifact of some implicit bias. In this paper we systematically study token-decoding explainability across diverse ViT settings, spanning architectures, scales, and training objectives, using both controlled experiments and real-world models. We find that while some models indeed exhibit aligned and interpretable token decodings, others exhibit either inverted explainability maps or a complete lack of correlation with the image semantics. As we show, these tendencies toward different explainability behaviors are not intrinsic to the architecture and can vary arbitrarily with changes in the optimization landscape. We provide theoretical support for these observations by demonstrating a setting where a single fixed toy transformer architecture can provably implement the optimal classification rule in multiple ways, which differ in their patch-decoding explainability patterns. Our results show that token-decoding explainability is not an inherent property of trained transformers, but rather an outcome contingent on the problem setting and thus should not be a-priori assumed to hold. Code can be found in https://ronensc.github.io/Explainability-Illusion/ .

0.18

Selling in the Wild: Benchmark and Randomized Field Evidence for a Sales Agent Trained on Commercial Conversations

cs.IR cs.CY ·2026-10-18 ·Benjamin Tannenbaum
benchmark
Abstract

We test whether Javo, a sales agent fine-tuned on 4.6 million consented, de-identified commercial conversations, sells better than a frontier general-purpose model and better than human sales associates. The evaluation has four parts. Study 1 scores the agent on two public sales benchmarks, SalesLLM and SmartSalesBench, against 10 frontier and open-weight models. Study 2 is a randomized controlled trial on 100 live websites in 14 industries (e-commerce, travel and hospitality, and high-consideration services), run from 1 May to 30 September 2026, in which 14.8 million visitors were randomized to no agent, Javo, or a frontier-model agent with an identical interface, catalog tools and guardrails. Study 3 isolates the effect of the model by comparing those two agent arms. Study 4 compares Javo with each site's own human associates on 20 sites with staffed live chat. Relative to no agent, Javo changed revenue per visitor by +10.0% (95% CI +5.1 to +14.9). Among engaged visitors, it changed conversion by +21.6% relative to the frontier-model agent (95% CI +16.9 to +26.3). Against human associates, its engaged conversion was 11.26% versus 11.20% (non-inferior, margin -7%, one-sided p = 0.004; no significant superiority), at 1.8% of the cost per conversation. Return rates, unsupported product claims and complaint rates met the prespecified safety guardrails.

0.18

Persistent Representational Misalignment Induces Vision-Language Model Failures

cs.CV cs.LG ·2026-10-18 ·Yonatan Gideoni, Yoav Gelberg, Tim G. J. Rudner, Yarin Gal
Abstract

Vision-language models (VLMs) fail at some tasks that are simple for humans---but why? We hypothesize that many such failures stem from VLMs' difficulties translating between vision and language representations, a misalignment the Platonic Representation Hypothesis predicts should disappear with scale. This paper is composed of two parts: in the first, we show that existing vision and language representations are linearly unalignable. This misalignment is an inherent property of the data and hence expected to persist regardless of either scale or the alignment's complexity. In the second, we demonstrate how this misalignment both leads to a predictably induced generalization failure and correlates with existing failure modes, where more representationally aligned VLM backbones accordingly exhibit fewer failures. Overall, these results imply representations will not become aligned simply given sufficient scale, providing evidence against a strong version of the Platonic Representation Hypothesis.

0.18

TIRE: Top-Down Formula Instantiation with Reliable Execution for Physics Problem Synthesis

cs.CL cs.AIcs.LG ·2026-10-18 ·Xiaoyu Wu, Zhaoxin Yu, Xinglin Xiao, Nan Xu
Abstract

Advancing physics reasoning in large language models requires scalable training data that connect physical laws, concrete problem scenarios, and verifiable numerical answers. Expert-designed problems offer top-down control over physical laws, conditions, and scenarios, but require costly human effort. LLM-based synthesis from texts scales efficiently, but model-generated answers and assessments leave supervision vulnerable to hallucinations. We introduce TIRE, a framework for Top-Down Formula Instantiation with Reliable Execution that brings knowledge-driven problem construction to large-scale data generation. Top-down instantiation uses reusable physical formulas and their applicability conditions to control the physical relations and numerical configurations of generated problems. Reliable execution establishes reference answers before problem generation, providing a common numerical foundation for problem construction and teacher distillation. Together, these mechanisms combine control over problem content with execution-grounded supervision for both supervised fine-tuning and reinforcement learning. Using TIRE, we construct TIRE-19k, spanning nine subfields of physics. Across 3 backbone LLMs and 4 benchmarks, TIRE-19k outperforms 5 representative datasets with equal amounts of SFT data, exceeding the strongest baseline by 2.86 percentage points on average. Further GRPO raises Qwen3-8B's average accuracy by 1.52 percentage points over its TIRE-19k SFT checkpoint. These results demonstrate the value of generating physics training data from explicit physical knowledge with reliable execution.

0.17

Shared-Private Cross-Modal Discrepancy Learning for Imbalanced Multi-Source Audio-Visual Deepfake Detection

cs.SD cs.AIeess.AS ·2026-10-18 ·Qiwu Wen, Jahangir Alam
Abstract

Audio-visual deepfake detectors are often trained under a single dataset or manipulation regime, but multi-source training collections may contain heterogeneous attacks and uneven source distributions. Training separate detectors for each source is inefficient, while a single shared detector may be biased toward dominant source-specific patterns. We study known-source imbalanced multi-source audio-visual deepfake detection using FakeAVCeleb and LAV-DF, where source identity is supplied during training and inference. To address this setting, we build on the Multi-Scale Cross-Modal Transformer Encoder (MSCT) and propose Pre-MSCT Shared-Private Feature Fusion. Our method routes \texttt{[CLS]}-augmented audio and visual tokens into a shared stream and source-specific private streams before the first MSCT block, while retaining its internal multi-scale self-attention (MSSA), differential cross-modal attention (DCA), and feed-forward operations. The shared stream models cross-source reusable evidence, while the private streams retain source-dependent residual variations. A domain-aware gated residual fusion module injects selected private evidence into the shared representation for the main fused prediction. Experiments under two controlled source-composition protocols show improved macro and worst-source AUC compared with the MRDF and MSCT baselines. These results support shared-private cross-modal discrepancy learning in the evaluated two-source setting with known source identity.

0.17

VoxDub: A Bilingual Lip-Synchronized Audio-Visual Dataset for Automatic Video Dubbing

cs.SD eess.AS ·2026-10-18 ·Kaidi Wang, Wenhao Guan, Peijie Chen, Weijie Wu, Xuetao Lin +2 more
Abstract

Automatic video dubbing (AVD) requires temporally aligned speech, transcripts, and video, yet existing corpora vary in audio quality, transcription reliability, and speaking-face visibility. We introduce VoxDub, a bilingual lip-synchronized audio-visual dataset containing 766K clips and approximately 1.58K hours of in-the-wild Chinese and English recordings, covering varied speakers, scenes, speaking styles, and linguistic content. Its curation pipeline combines speech-centered segmentation with dual-ASR transcription checking, speech quality filtering, face visibility checks, and lip-speech synchronization verification. Text-to-speech (TTS) experiments show that training on VoxDub yields lower word error rates and higher speaker similarity than training on the compared audio-visual corpora, LRS3 and CN-CVS. AVD experiments show broadly comparable speech quality and subjective synchronization. These results support the utility of VoxDub for TTS and AVD training. More details are available on the project website (https://voxdub-dataset.github.io/VoxDub/).

0.17

Residential Construction Image Dataset Using CSI Divisions for Deep Learning-based Object Detection

cs.CV cs.AI ·2026-10-18 ·Fahad Iqbal, Shayan Mirzabeigi
benchmarkdataset
Abstract

Accurate recognition of construction site elements is essential for the application of deep learning (DL) in the construction industry. Despite rapid advances in computer vision, the adoption of DL in construction has been limited by the lack of structured, publicly available datasets aligned with established industry standards. To address this gap, we introduce the Residential Construction Dataset Using CSI Divisions (ResCD-CSI) image dataset, a manually labeled image benchmark organized according to the divisional structure defined by the Construction Specifications Institute (CSI). The dataset comprises high-resolution images collected from active residential construction sites across the United States, capturing diverse building components, materials, and equipment under real-world conditions. In total, 16,365 objects spanning 14 major CSI divisions were precisely annotated using bounding boxes, ensuring consistency and reliability for object detection tasks. By aligning with the CSI divisional framework, ResCD-CSI establishes a standardized resource that bridges construction practice with computer vision research and supports the development of Artificial Intelligence-driven solutions for residential construction monitoring.

0.17

Design Considerations in Low-Resource Archival OCR: A Case Study in Lakota

cs.CL ·2026-10-18 ·Téa Wright, Laura X. Ma, Claudia Iron Hawk, Hannah J. Haynie, Alexis Palmer +3 more
Abstract

Building and expanding access to digital archives supports language reclamation and creates lasting resources for community revitalization. Designing effective tools to support these efforts requires taking into account the community-, language-, and document-specific context. We investigate how small ($\leq$3B) vision-language models (VLMs) can be adapted as optical character recognition (OCR) tools for Lakota, a North American Indigenous language. In addition to benchmarking modeling approaches, we inspect four practical facets of low-resource archival OCR that arise in previous work and conversations with community members: constrained training budgets, segmentation noise, scan degradation, and orthographic variance. We demonstrate that fine-tuning VLMs for OCR post-correction on 35 pages of transcribed data decreases character error rate (CER) from 17.1 to 3.9\% on a bilingual corpus, and find that training only with nine pages increases CER by only 1.1\% over this. However, evaluating under realistic automatic segmentation nearly doubles the CER of the best-performing models, stressing the importance of taking realistic use cases into account. We hope this work encourages future research on developing language technologies grounded in the workflows and constraints of ongoing revitalization efforts.

0.17

Behave Like Experts: Fusing Data with Semantic Knowledge for Effective and Efficient Semantic Map Construction

cs.CL ·2026-10-18 ·Zhu Liu, Zhen Hu, Yuxuan Liu, Ying Liu
benchmark
Abstract

Semantic map models (SMMs) construct a graph that reflects conceptual similarity by comparing cross-lingual form-meaning mappings. Experts build such models not only by obeying the data constraint, i.e., the connectivity hypothesis, which states that conceptual nodes belonging to the same form must be connected, but also by incorporating their prior semantic knowledge. Current automatic methods ignore this knowledge, which harms both accuracy and efficiency by enlarging the search space of candidate graphs. In this paper, we fuse the distributional semantic knowledge hidden in the internal representations of language models into a top-down, data-driven algorithm. This expert-like behavior re-sorts the candidates in a huge search space, guided by conceptual semantics similarity. Experiments on a self-collected benchmark of 10 datasets and comprehensive metrics show that our method builds SMMs both effectively and efficiently.

0.16

DiaLong: Benchmarking Reminiscence in Long-Context Dialogues

cs.CL cs.AI ·2026-10-18 ·Owen Bianchi, Samer Aslan, Ryunosuke Saito
benchmarkdataset
Abstract · Equal contribution. Work conducted in 2023 at Johns Hopkins University

How well can transformer-based Large Language Models (LLMs) retain and recall memories in long contexts of dialogue? LLMs are known to forget long-range factual information in their context, yet the field lacks a comprehensive benchmark to quantitatively assess these shortcomings. We present DiaLong, a long-context dialogue dataset and accompanying memory task that standardizes evaluation in this space. It consists of 100 unique conversations created by humans and then augmented with LLMs, extending beyond lengths found in existing datasets. Our selected memory task is fact recall: classify 40 candidate facts about a dialogue as true or false. This task directly probes the model's ability to actively retain and retrieve information, serving as a critical measure of memory within extended conversational contexts. In testing both the latest LLMs and proposed memory enhancements on DiaLong, we demonstrate that more work needs to be done to enable LLMs with better memory performance. By enhancing memory capabilities, LLMs can improve in maintaining coherence, relevance, and human-like interaction over extended conversation. Improving these capabilities will help LLMs fulfill diverse user needs and expectations in various conversational scenarios.

0.16

Learning Speaker-Activity-Conditioned WavLM Speaker Embeddings with Real Conversational Data

eess.AS ·2026-10-18 ·Prachi Singh, Nguyen Hai Phong, Petr Pálka, Jiangyu Han, Marc Delcroix +3 more
published
Abstract · Accepted to SLT 2026

End-to-end neural diarization with vector clustering (EEND-VC) is a strong framework for long-form conversational diarization, where speaker embeddings are used to link local speaker decisions across windows. Conventional EEND-VC systems extract these embeddings by cutting out regions assigned to each local speaker, preferably using only non-overlapped speech. This works well when enough clean target-speaker speech is available, but it discards useful speaker evidence in overlapped regions and becomes unreliable when a local speaker appears mostly in overlap. We propose a speaker-activity-conditioned embedding extractor that adapts a pretrained WavLM-MHFA speaker verification model to extract target-speaker embeddings directly from full conversational windows. Inspired by the recently proposed diarization-conditioned Whisper (DiCoW) model for target-speaker ASR, frame-level diarization-dependent transformations inject local speaker-activity information into intermediate WavLM representations while preserving the pretrained initialization. Experiments on multiple conversational diarization benchmarks show that fine-tuning with limited real meeting data reduces speaker-confusion errors and improves EEND-VC diarization performance. The results further indicate that real conversational data provides an effective alternative to large-scale simulated-mixture training for diarization-aware speaker embedding extraction.

0.16

Modaliff: A Unified Token Diffusion Model for Any-to-Any Generation

cs.CV cs.AI ·2026-10-18 ·Duc Kieu, Kien Do, Minh Khoa Le, Tung Kieu, Dang Nguyen +1 more
Abstract

Any-to-any generation seeks a single model that translates flexibly among text, image, and audio, yet existing approaches either require fully aligned multimodal tuples or a shared representation that suppresses modality-specific structure. We introduce Modaliff, a novel token-based diffusion model that learns from widely available bimodal pairs while preserving each modality in its native latent sequence. Modaliff assembles conditions and targets into a unified token sequence, with per-token attributes to disambiguate token identities and role-based positional embedding to preserve each modality's internal structure across sequence layouts. This design enables a single denoiser to support diverse generation tasks without task-specific architectural changes or a shared cross-modal latent space. Beyond one-to-one generation, Modaliff supports multiple conditions either by combining single-condition predictions at inference or by lightweight finetuning on captions that jointly describe multiple inputs. Experiments across text, image, and audio show that Modaliff achieves the best fidelity among any-to-any models on most generation tasks while using substantially fewer parameters and training data.

0.16

Visual Scaffolding: Toward Native Multimodal Reasoning

cs.CV ·2026-10-18 ·Run Shao, Zhaoyang Zhang, Ziyu Li, Linrui Xu, Ji Qi +4 more
Abstract

The success of chain-of-thought (CoT) in complex reasoning can be attributed to a textual scaffolding effect: explicit intermediate results can be examined, reused, and composed to support subsequent reasoning, much as scaffolding supports a building under construction. However, mainstream multimodal LLMs lack an equivalent scaffold for visual distinctions that are difficult to express in concise verbal descriptions. We therefore ask: Can native visual representations become first-class citizens alongside language, providing a visual scaffold for reasoning? We define native multimodal reasoning as reasoning in which native visual representations and language jointly carry reusable intermediate states. As a concrete implementation, the visual scaffolding mechanism we proposed lets the MLLM autonomously initiate a retrieval action at any reasoning step. A lightweight retriever selects native visual features, which are then incorporated directly into the evolving CoT and remain available to subsequent steps. To evaluate this mechanism under controlled visual demands, we introduce NMR-Bench, comprising Burden Counting and Path Tracing. With 88M added parameters (approximately 2%) on Qwen3-VL-4B-Thinking, our model achieves 66.43% and 72.50% accuracy, respectively, surpassing the evaluated GPT-5.5 and Gemini 3.1 Pro baselines. Controlled interventions show that changing selected visual content affects task accuracy and can redirect subsequent visual selection even when intervening text is held fixed. These findings show that native visual states can function as intermediate reasoning material and point toward reusable scaffolds for additional modalities.

0.15

When Semantics Fall Short: Persona-Guided Cross-Domain Sequential Recommendation with Large Language Models

cs.IR ·2026-10-18 ·Jundong Chen, Honglei Zhang, Shuai An, Enneng Yang, Han Yu
Abstract

Cross-domain sequential recommendation (CDSR) alleviates data sparsity by transferring user behaviors across domains. Recent advances increasingly leverage large language models (LLMs) to bridge domains through global item semantics. However, this paradigm is inherently limited by two issues: I. semantic domain isolation, where item semantics remain domain-specific and hinder transferable modeling, and II. semantic-behavior inconsistency, where local semantic relations are more consistent with behavioral transitions, whereas such consistency often breaks down globally across domains. To this end, we propose PerCDR, an LLM-enhanced CDSR framework that shifts the focus from global semantic bridging to persona-guided behavioral modeling. Specifically, PerCDR leverages LLM-inferred personas to globally guide transferable cross-domain behavioral patterns, while exploiting local semantic relations to enhance intra-domain behavioral modeling, followed by domain preference adaptation for recommendation refinement. Extensive experiments demonstrate the superiority of PerCDR and validate its effectiveness.

0.14

Sufficient Frame Subsets for Video Understanding

cs.CV ·2026-10-18 ·Lahiru Wijayasingha
Abstract

Understanding when during a video a model makes its critical observations is essential for interpreting video-understanding systems. However, existing attribution methods often spread evidence diffusely across near-duplicate frames. We introduce Sufficient Frame Subsets: compact sets of frames in which all other frames are masked, yet each subset independently preserves the model's prediction. We evaluate our method on two complementary benchmarks, UCF101 and Something-Something v2, using both CNN- and transformer-based architectures. Our results show that the proposed algorithm identifies highly compact and effective frame subsets for explaining model decisions. We also find that models capable of relying on a greater number of sufficient frame subsets tend to perform better, possibly because they are more robust, consistent with prior research. Finally, the extracted frame subsets help us analyze model failure cases, reveal which aspects of a video the model has learned to emphasize, and identify opportunities for improving future training procedures.

0.14

Map2Vid: Embodied Navigation Video Generation with Human-Friendly Motion Control

cs.CV ·2026-10-18 ·Yulin Dai, Yuqian Fu, Tianwen Qian, Zhen Xing, Bojia Zi +1 more
benchmarkdataset
Abstract

Controllable video generation has recently achieved impressive visual quality, yet existing control paradigms remain poorly suited to embodied navigation. Text prompts are spatially ambiguous, while camera-trajectory control relies on low-level motion specifications that are unintuitive for human users and difficult to manipulate. In this work, we introduce Map2Vid, a novel task that reformulates embodied navigation video generation as a BEV-map-to-video synthesis problem, together with a large-scale benchmark for systematic evaluation. By using BEV maps, which provide a visual representation of global scene layout and intended navigation routes, Map2Vid enables human-friendly navigation video generation. To support this task, we construct a dedicated dataset with over 100K navigation sequences across 61 indoor scenes, each paired with egocentric videos, BEV maps, camera trajectories and language instructions. We further develop a unified framework that integrates multiple control signals, together with a curriculum training strategy that progressively weakens non-map inputs, enabling flexible inference under both full conditioning and map-only input. Extensive experiments across multiple foundation generation models demonstrate that controllable navigation video generation via BEV map is a viable and promising direction.

0.13

ViMoLa: Discrete Motion Tokens for Language-Guided Video Generation

cs.CV ·2026-10-18 ·Bishoy Galoaa, Sarah Ostadabbas
Abstract

Language-guided video generation often captures scene content while missing the requested motion. We present ViMoLa (Video Motion Language), which predicts a motion prior from language and injects it into a frozen image-to-video backbone. A spatial vector-quantized autoencoder compresses dense trajectories into 64 tokens from a 512-entry codebook; a 28M-parameter transformer predicts the tokens from text; and zero-gated cross-attention adapters condition every frozen Wan~2.2 DiT block through forward hooks. Trained on 462K videos from Something-Something~v2 and Kinetics-400, ViMoLa raises directional accuracy from 0.287 to 0.442 (chance 0.250), with similar measured generative quality. Across-prompt shuffled codes achieve 0.283, supporting the importance of prompt-matched token content. A low-rank adapter reaches 0.428 at 1.45% additional forward-pass FLOPs, and reusing the tokenizer and language model with newly trained CogVideoX adapters raises accuracy from 0.301 to 0.398. Control remains imperfect: oracle codes reach 0.531, compound-motion gains are smaller, and the adapter does not explicitly preserve token-to-patch assignments.

0.13

MotionShift: Environment-Compatible Motion Transfer via Planner-Evaluator Synergistic Adaptation

cs.CV ·2026-10-18 ·Yanchen Liu, Junyao Gao, Yanan Sun, Wensheng Chen, Guangming Lu +1 more
Abstract

Existing motion transfer methods excel at exact, rigid motion replication from reference videos to target scenes, whereas they often fail to account for the compatibility between the transferred motion and the target scene environment in accordance with physical laws. In this work, we propose MotionShift, a novel motion transfer framework that achieves high-fidelity motion transfer while ensuring physical compatibility with the target scene. In particular, we design a planner--evaluator iterative motion adaptation mechanism, in which a Large Language Model (LLM) acts as the planner to generate motion trajectories by leveraging its strong semantic reasoning capabilities, while a Vision-Language Model (VLM) serves as the evaluator to rectify the motion based on its superior visual perception and scene understanding. Through this planner--evaluator synergistic reasoning process, the motion is adapted to the target scene in a training-free manner. The adapted motion is then used as a generative condition for an image-to-video (I2V) model to render videos in the target scene. Extensive experiments demonstrate that MotionShift achieves high-fidelity and physically compatible motion transfer across diverse and complex scenes. Moreover, we show that MotionShift can be readily integrated into existing trajectory-conditioned I2V models, endowing them with realistic motion transfer capabilities.

0.13

Self-Distillation for Supervised Fine-Tuning

cs.CL ·2026-10-18 ·Wenxuan Zhu, Bang An, Shuming Liu, Wenqing Cui, Tianqi Xu +4 more
Abstract

Supervised fine-tuning treats the expert demonstration or gold response as a one-hot target, suppressing valid alternative continuations that the model would otherwise assign probability to. We propose Gold-Conditioned Distillation (GCD), a method that turns expert demonstrations into model-native supervision. GCD uses one model as both teacher and student: the teacher conditions on a prompt containing the gold response in context, the student conditions on the original prompt as at inference, and the student is trained to minimize the forward KL divergence to the teacher's next-token distribution at each gold response token. This lets the model use its prior knowledge to teach itself the new task: correctness is anchored by the gold response, while the soft target preserves plausible alternatives. Across various training tasks on Qwen and Llama models, GCD improves new-task accuracy over SFT and generally reduces degradation on general benchmarks.

0.11

PCXR-Copilot: Device-aware pediatric chest radiographic analysis with segmentation, suppression, and edema-related cues

cs.CV ·2026-10-18 ·Xueyang Li, Alec Girten, Peng Zeng, Mingze Jiang, Shaoyang Huang +4 more
Abstract

Patients admitted to pediatric ICU and cardiac ICU often require intensive perioperative support and frequent chest radiography to monitor support devices, extravascular lung water, and body-wall edema. Radiographic interpretation and analysis is often difficult due to superimposed anatomic and iatrogenic structures that obscure clinically relevant signals. While deep learning can provide decision support, translation to pediatric radiographs is still limited by high device burden and scarcity of pediatric datasets with clinically grounded device annotations. To address this gap, we first introduce PediaDeviceSeg-1K, a curated dataset of 1,000 anterior-posterior radiographs of pediatric CICU patients with clinically grounded, fine-grained annotations of common ICU devices. To our best knowledge, this is the first publicly available pediatric CICU CXR resource with clinically grounded device annotations for device-focused benchmarking. Building on PediaDeviceSeg-1K, we further present PCXR-Copilot, a device-aware pipeline that outputs lung and device masks, optional device-suppressed views, and edema-related visualization maps and scalar indices. The full pipeline has been integrated and runs in real time within the TWIST platform at Boston Children's Hospital, demonstrating feasibility for real-world clinical deployment.

0.11

Towards Figurative Language Understanding in Moroccan Darija: An Evaluation of Large Language Models

cs.CL ·2026-10-18 ·Jihad R'baiti, Youssef Hmamouche, Emmanuele Chersoni, Chu-Ren Huang, Amal El Fallah Seghrouchni
benchmark
Abstract

Figurative language remains a major challenge for Machine Translation (MT) systems, as the intended meaning of such expressions often goes beyond literal translation. Although recent advances in Large Language Models (LLMs) improved the translation of Arabic dialects, benchmark resources for evaluating figurative language translation in Moroccan Darija remain scarce. To this end, we introduce the Moroccan Darija Figurative Language Corpus (MDFC), a benchmark dataset for MT and multiple-choice question (MCQ) selection covering idioms, metaphors, and proverbs. Using MDFC, we evaluate 19 open- and closed-weight language models and MT systems at both the expression and contextual levels under zero-shot and six-shot settings. We also conduct layer-wise similarity-based probing to determine how model representations distinguish figurative from literal meanings. Our findings reveal that closed-weight models achieve the strongest translation performance, while the benefits of contextual information vary across models and tasks. Differences between translation performance and MCQ accuracy highlight the distinction between generating an appropriate figurative translation and recognizing the intended meaning among candidate translations.

0.08

IECLA: Interpretable Emotion Contribution Learning for Affective Image Editing

cs.CV cs.AI ·2026-10-18 ·Yijie Guo, Dexiang Hong, Weidong Chen, Xinyan Liu, Zhendong Mao
Abstract · 31 pages, 10 figures, 24 tables; under review at ICLR 2027

Affective image editing (AIE) aims to evoke a target emotion while preserving the content and structure of a source image. Existing methods improve emotional expression or visual fidelity, but the target emotion alone does not reveal which source regions should be changed and which should remain intact, especially in complex scenes. We introduce Interpretable Emotion Contribution Learning for Affective Image Editing (IECLA), a framework that separates source-emotion attribution from target-aware editing. IECLA constructs signed object-level supervision through counterfactual interventions and trains a context-conditioned contribution predictor to localize evidence that supports or opposes the source emotion. At inference, an image-level evaluator estimates the source emotion directly from the input, while the predictor identifies editable and preservation evidence without repeated interventions. A target-aware editing agent then translates this evidence and the desired emotion into localized, content-aware operations for existing editors. Extensive evaluations show that IECLA predicts counterfactual contributions more faithfully on both in-domain and cross-dataset tests, while external VLM and independent human references support the quality of its localized evidence. Across six editing backbones, IECLA consistently strengthens target-emotion expression while preserving image content and structure, and user studies further show improved perceived target expression and naturalness.

0.08

Tricking Synthetic Image Detectors with Amortized Generative Attacks

cs.CV ·2026-10-18 ·Mohamed Awad, Khaled Elhady, Yasser Rohaim, Samuele Poppi, Nils Lukas
Abstract

Rapid advances in image synthesis have increased the need for reliable synthetic-image detectors, yet their robustness to transferable attacks remains incomplete and understudied. Existing evaluations generally examine whether an attack constructed against one detector transfers to another, but rarely isolate how its effectiveness changes when the attacked images originate from different generators. We conduct a systematic evaluation of three representative attacks: two iterative perturbation-based methods, MIG and MuMoDIG, and the amortized generative attack Tricker. Across four detectors and five generators spanning diverse architectures, we introduce a two-axis evaluation protocol that independently varies the surrogate detector used to construct or train the attack and the generator supplying its input images. We assess attack effectiveness, cross-detector transferability, perceptual quality, runtime, and interaction with common post-processing operations. We revisit Tricker and show that its data and training requirements can be substantially reduced, enabling broader and more efficient evaluation. Our results demonstrate that transferability is not a detector-only property: the input-generator distribution strongly influences both evasion effectiveness and attacked-image quality. MIG and MuMoDIG exhibit pronounced dependence on the detector--generator pair, whereas Tricker provides more uniform white-box evasion and image quality. We additionally observe that post-processing maintains the asymmetric transferability topology. These findings expose a neglected dimension of attack evaluation and motivate transfer-aware protocols that jointly account for detector and generator variation. We will release our code, evaluation data, and trained attack models with a responsible-use statement.

0.08

CASE: Complementary Articulation-Aware Speech Encoder to Enhance Speech-Driven 3D Facial Animation

cs.CV cs.SDeess.AS ·2026-10-18 ·Jaehong Myung, Seungho Eum, Taegon Kim, Hyunjung Chung, Unsang Park
Abstract

General-purpose self-supervised speech encoders provide strong acoustic representations, but their pretraining objectives are not explicitly aligned with the visible articulatory dynamics required by speech-driven 3D facial animation. We propose CASE (Complementary Articulation-aware Speech Encoder), a complementary speech encoder adapted by contrastively aligning short speech clips with continuous 3D lip-mesh VQ-VAE latents. After adaptation, the geometry branch is removed, and the audio-only CASE representation is fused with the baseline SSL speech representation through residual cross-attention, while retaining the motion-decoder architecture and objective. Across four VOCASET baselines, CASE reduces mean lip vertex error by 2.8-6.0%, and mean full-face vertex error decreases in every system. Under the same fusion protocol, CASE yields lower mean LVE than the unadapted AV-SE stream on all four generators, with model-dependent margins. Because adaptation uses Korean speech-geometry pairs and evaluation uses English VOCASET, the result is consistent with cross-language downstream transfer under a shared English-pretrained audio-visual prior.

0.07

Security & Privacy 18

GreenNIDS: A Benchmark Workflow and Analysis Dataset for NetFlow Intrusion Detection Experiments

cs.CR cs.LG ·2026-10-18 ·Lilia-Eliana Popescu-Apreutesei, Sabina-Cristiana Necula
publishedbenchmark
Abstract · 12 pages, 3 figures, 2 tables. Accepted at IE 2026 (International Conference on Informatics in Economy)

Comparing the predictive performance and computational cost of flow-based intrusion detection experiments requires both quantities to be available for each run. GreenNIDS connects benchmark execution on public NetFlow datasets to a reusable analysis dataset. Model implementations write a common set of outputs, which are ingested into a run-level table with companion tables for model configuration, feature importance, and training history. The run-level table combines predictive metrics with timing, throughput, memory use, and energy estimates based on thermal design power. A CPU-only demonstration comprises 216 runs across four dataset families, three versions, and binary and multiclass tasks. Ingestion produces a 216 x 96 analysis table with no duplicate run identifiers and complete Matthews correlation coefficient (MCC) and F1 coverage. Multiclass ROC-AUC is missing for 36 PyTorch runs, making metric availability an explicit part of subsequent comparisons. The stored records support grouped performance and training-time plots, as well as resource screening after execution. Combining task-specific MCC thresholds with limits on training time, peak RAM, and estimated training energy retains 25.0% of the full run collection. These results establish compatibility within the evaluated collection and support comparisons after execution, subject to the recorded preprocessing, execution profile, and approximate energy accounting.

0.49

How Much Feedback Is Needed to Break Template Protection? The Geometric Vulnerability of Biometric Authentication

cs.CR ·2026-10-18 ·Eliron Rahimi, Margarita Osadchy, Orr Dunkelman
Abstract · preprint

Biometric authentication systems increasingly use privacy-preserving mechanisms to protect sensitive templates. However, the authentication interface itself may still leak information that enables reconstruction attacks. Prior attacks exploit released similarity scores, whereas more advanced systems reveal only the binary accept/reject decision. This binary feedback provides substantially less information, making template reconstruction far more challenging. In this work, we introduce query-efficient reconstruction attacks under extremely limited feedback, addressing two key limitations of prior attacks. First, we introduce the first attack that reconstructs protected biometric templates using only authentication success/failure feedback by exploiting the geometric constraints of biometric authentication. Second, we introduce distribution-aware query-efficient reconstruction methods for score-releasing systems, achieving a more realistic threat model by reducing the number of required queries by a factor of $5\times$ compared to recent prior score-based attacks. Finally, we combine these two insights by incorporating a pretrained facial diffusion prior into our geometric binary attack, reducing the number of required queries by a factor of $5\times$. Our findings reveal a fundamental security limitation of biometric authentication: independently of the template protection mechanism, the authentication process itself exposes geometric information that can enable perfect template reconstruction up to numerical precision. Moreover, the distribution of biometric templates can be leveraged to make significantly more effective reconstruction attacks.

0.43

SoK: PIAtlas: Decomposing the Unbounded Prompt Injection Attack Space

cs.CR cs.CLcs.LG ·2026-10-18 ·Stavros Damianakis, Konstantinos Psounis
surveybenchmarkcode/data
Abstract · 61 pages, 1 figure, 23 tables. Code and data: https://github.com/stevendamianakis/piatlas | Interactive taxonomy: https://stevendamianakis.g…

Prompt injection is the dominant security risk for large-language-model (LLM) applications and agents, yet the field still lacks a coherent map of how these attacks are constructed. More than twenty public datasets describe prompt injection using incompatible vocabularies, often organized around delivery vector, attacker goal, or generation method. In this Systematization of Knowledge, we present PIAtlas, a taxonomy organized instead around how the injected instruction is constructed to subvert the underlying model. Its six classes, one per subversion mechanism, form the roots of a hierarchy of techniques and sub-techniques that researchers can use to describe attacks consistently and design more systematic evaluations. We derive the taxonomy from how LLMs are trained, prompted, and influenced by human-like framing, and compare it with historical attack data. To ground this comparison, we release PICorpus, to our knowledge the largest license-clean collection of public prompt-injection datasets, and map thousands of its records to PIAtlas under a human-verified protocol. The mapping shows that attacks usually combine several techniques drawn from only a few mechanisms. It also reveals gaps that existing labels obscure: widely used agent benchmarks test only part of the attack space, and adaptive attackers favor mechanisms that curated benchmarks underrepresent. PIAtlas thus provides a common foundation for understanding this attack space, improving benchmark coverage, and guiding red teaming and defenses.

0.32

Evidence-Grounded LLM-Assisted Android Mobile Forensic Reporting

cs.CR cs.CY ·2026-10-18 ·Madhab Chandra Das, Sajid Ahmed, Tasnia Tabassum, Qingzhong Liu
published
Abstract · 10 pages, 3 figures, 7 tables. Accepted at the 8th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems, and …

Large language models can assist forensic report drafting, but fluent output does not ensure evidentiary reliability. This study presents an evidence-grounded framework for evaluating LLM-assisted Android forensic reports at the atomic-claim level. Gemini, Llama, and Qwen each generated reports for two controlled cases: a public Android 13 dataset and a physical LG device with a zero-reportable-fact boundary. The six accepted reports were decomposed into 612 claims and evaluated against frozen, human-reviewed authorities through content labeling, citation validation, fact-recall analysis, independent rechecking, and dispute resolution. The final corpus contained 424 supported, 163 unsupported, four contradicted, four overstated or incomplete, and 17 excluded claims. In the public case, all three models recovered all 27 authorized report-level facts. Citation validity nevertheless differed substantially: Gemini produced 168 valid citations from 189 required citations, Qwen produced 145 from 189, and Llama produced none from 135. In the physical device case, the models generated 65 claims despite the zero reportable-fact boundary, including 38 unsupported and one contradicted claim. These results show that evidence restriction and complete fact recall do not guarantee claim support or valid evidence attribution. The framework provides an auditable method for detecting unsupported content, preserving claim-to evidence lineage, and retaining investigator authority over final forensic reports.

0.32

Federated Learning for Malware Image Classification under Data Heterogeneity

cs.CR cs.DCcs.LG ·2026-10-18 ·Victor Taiwo, Cemal Nisan, Muhammad Athallah, M. Emre Gursoy, Oznur Ozkasap
published
Abstract

Malware image classification (MIC), in which binary files are converted into visual representations and deep neural networks are trained to identify malware families, has recently emerged as an effective approach for malware detection. Federated learning (FL) enables malware image classifiers to be trained collaboratively by multiple clients without sharing raw data. However, a fundamental challenge in FL is data heterogeneity (non-IID), which is particularly relevant but understudied in the malware context. In this paper, we address this gap via three contributions. First, we present an empirical study of federated MIC under varying degrees of data heterogeneity across three popular malware image datasets (MalImg, Virus-MNIST, and MalNet). We show that while FL achieves strong accuracy under IID conditions, both accuracy and training stability degrade as data becomes increasingly non-IID. Second, we introduce a set of round-level statistics that characterize the data contributed by participating clients in each FL round, and demonstrate that class entropy (diversity of class labels present in a given round) exhibits strong positive correlation with global model accuracy. Third, leveraging this insight, we propose an entropy-aware client selection strategy that filters candidate client subsets based on class entropy, and show that our strategy outperforms standard random selection. Our results highlight the role of client composition in federated MIC and underscore the importance of distribution-aware classifier training under data heterogeneity.

0.29

AdaFedXplain: Adaptive Federated Learning with Explainable AI for Ransomware Detection

cs.CR cs.AI ·2026-10-18 ·Md. Emran Biswas, Md. Delowar Hossain, Mst. Tanbin Yasmin Tanny, Muhammad Aminur Rahaman, Md Al Amin +1 more
Abstract · https://nsclab.org/ml4cs2026/ (Eighth International Conference on Machine Learning for Cyber Security (ML4CS 2026) )

Attacks using ransomware have developed into complex multiphase campaigns that target vital infrastructure, resulting in annual damages of billions of dollars. Conventional detection techniques are ineffective against zero-day variations, and privacy issues are raised by centralized machine learning algorithms. Federated learning allows for cooperative model training without data sharing, however current FL approaches suffer from explainability issues, high communication cost, and non-IID data heterogeneity. In this study, we present AdaFedXplain, an explainable AI-based adaptive federated learning system for ransomware detection. Our approach includes progressive XGBoost complexity to avoid overfitting, a decaying learning rate schedule for stable convergence, and dynamic client grading based on ransomware prevalence and data sufficiency. Using 62,485 real-world ransomware samples from four OS-based non-IID clients, we compare AdaFedXplain against FedAvg, FedProx, and SCAFFOLD. According to experimental data, AdaFedXplain outperforms all baselines with 99.59% detection accuracy and 99.64% F1-score while cutting energy usage by 75%, latency by 65%, and communication overhead by 75%. The ablation investigation demonstrates that every element enhances performance. Security analysts can use SHAP explainability to identify important PE header characteristics that influence detection decisions. Our system provides a collaborative ransomware detection solution that is efficient and interpretable, while reducing privacy risks through federated learnings data-local training paradigm.

0.28

SemVideoSte: Agentic AI for Cross-Modal Semantic Video Steganography Communication

cs.CR eess.SP ·2026-10-18 ·Rui Meng, Song Gao, Tingting Zhu, Zixun Tang, Jinke Ren +4 more
code/data
Abstract · 13 pages, 11 figures

We propose an agentic artificial intelligence (AI)-driven cross-modal semantic video steganography communication (SemVideoSte) scheme for secure cross-modal information transmission. Agentic AI interprets the service context to determine a semantically matched public prompt and coordinates video generation, message embedding, semantic transmission, and message recovery process. A multi-mask mechanism combines variational autoencoder (VAE) reconstruction robustness, semantic transmission robustness, and latent separability to select reliable embedding positions. Meanwhile, a private digital token controls the random seed and guidance offset to construct positive and negative terminal latent states, enabling coverless message embedding during the final diffusion sampling step. Simulation results on open-source datasets demonstrate that SemVideoSte achieves high embedding capacity, reliable message recovery, and strong semantic consistency under different channel conditions. In addition, eavesdroppers without the correct digital token cannot reliably recover the secret message, and the generated stego videos effectively resist the steganalysis detection. We further establish a prototype testbed to verify the practical feasibility of SemVideoSte.

0.27

Towards End-to-End Privacy-Preserving AI for Real-Time ECG Monitoring

cs.CR ·2026-10-18 ·Beyazit Bestami Yuksel
published
Abstract · 9 pages, 4 figures, 2 tables. Accepted to MICCAI 2026 Doctoral Symposium (Springer LNCS)

Healthcare artificial intelligence is shifting from centralized archives toward distributed pipelines that combine wearables, edge processing, cloud analytics, and multi-institutional learning. In this regime, electrocardiogram (ECG) streams cross multiple trust boundaries, so privacy cannot be reduced to a single transmission control. Existing privacy-enhancing technologies typically secure isolated stages and leave gaps between acquisition, analysis, and institutional collaboration. This doctoral research develops a unified lifecycle framework for end-to-end privacy-preserving ECG AI. Rather than introducing another privacy mechanism, it orchestrates complementary privacy-enhancing technologies across the clinical ECG lifecycle, allocating each mechanism to the trust boundary where its security, computational, and operational properties are most appropriate. Three interoperable modules instantiate the framework: AES with CKKS-based homomorphic statistics for secure monitoring; adaptive chaotic protection for low-latency edge transmission; and federated learning with homomorphic aggregation under heterogeneous partitions. Across prototype monitoring, signal-security analysis, and simulated collaboration, the programme shows that privacy layers can be composed without collapsing diagnostic utility, while exposing overhead and fidelity limits that constrain clinical translation.

0.26

Adaptive Cyber-Defense for Evolving Attacker Capabilities using Reinforcement Learning and Behavior Trees

cs.CR ·2026-10-18 ·Nicholas Potteiger, Ankita Samaddar, Xenofon Koutsoukos
benchmark
Abstract

Autonomous cyber-defense agents trained using reinforcement learning (RL) are increasingly adopted for defending enterprise computer networks. However, existing approaches may become brittle as attacker capabilities evolve and enable new attack strategies. In this work, we study the problem of adapting defender policies learned in CAGE Challenge Scenario 2 (CC2), a benchmark enterprise network defense scenario, to an extended network scenario with new attacker capabilities represented by new attacker actions. Effective attack strategies that leverage new actions are initially unknown, and therefore attacker policies must be learned before defender adaptation can occur. To address this challenge, we develop adversarial RL (ARL) with behavior trees (ARLBT), which uses reward behavior trees (RBTs) to define progression-based reward functions and condition attacker policies on RBT execution. The proposed RBT incorporates the new attacker actions into structured attack strategies while leveraging parallel execution to generate rewards across multiple network nodes simultaneously. We evaluate ARLBT for learning attacker policies against two baseline defender policies across network scenarios with different sets of new attacker actions and compare against multiple attacker baselines. The results demonstrate that ARLBT improves attacker success compared to ARL using the original CC2 reward function and that conditioning policies on RBT execution can improve sample efficiency. We further show that defender policies can be adapted against the learned attackers, improving robustness to attacks that leverage new attacker actions.

0.25

Locate, Then Reason: End-to-End Patch Presence Testing in Stripped Binaries

cs.CR ·2026-10-18 ·Huaijin Wang, Tongda Wang, Zhibo Liu, Yue Zhang, Shuai Wang +4 more
Abstract

Binary patch presence testing (PPT) determines whether a security fix is present in a stripped binary. Prior systems often assume that patch-relevant code is known or successfully matched. Function-level matching is brittle to compilation changes, inlining, finite model context, and missing usage context; moreover, similarity suppresses the patch-sized distinction needed for the verdict. We present PatchAlign, an end-to-end binary PPT workflow, which performs large-scale localization that maximizes recall economically, then the final verdict reasons for patch-specific evidence. PatchAlign uses instruction-aligned retrieval and caller-callee context to accurately locate patch-relevant binary code pieces, then screens the shortlist and compares target binary code with vulnerable and patched references using an LLM. Our experiments indicate that PatchAlign can precisely localize patch-relevant code and determine patch status without a target-location oracle.

0.25

Don't Judge an Agent by Its Messages: Private-State Leakage under Passive Monitoring of Agent Communication

cs.CR ·2026-10-18 ·Daksh Pandey, Sayon Duttagupta, Xinhai Zou, Amit Singh Bhati
theory
Abstract

An AI agent can reproduce the distribution of its messages and of the receiver's task outcomes while disclosing private information that the task never required. We ask what a passive audit can certify when the auditor knows the authorised sender and the exact observable law but has no trusted pairing of messages with the private states behind them. For a single round, compatible mutual-information leakage forms a sharp interval, and fixing the entire task-relevant message law, with every declared task utility, can still leave a full private bit undetermined; total leakage is determined if and only if, in every task state, the remaining private state or the authorised message is a deterministic function of that state. For a sender, such as an LLM, that selects among its own responses, we prove that in a balanced case its change to the message law within each task state equals $(1-2^{-k})$ times the honest messages' dependence on the attribute, which also sets, up to a constant, how many messages a monitor needs. With Qwen2.5-1.5B-Instruct on a synthetic routing task, selection among 16 responses lets a public decoder recover a private bit in all 8,192 held-out rounds at 98.60\% routing accuracy (honest sender 98.90\%). A monitor that knows the selection rule reaches AUC 0.645 on single messages, close to the ceiling of 0.671 that honest messages predict, and 0.972 over 16, while generic monitors stay near chance. A law-preserving sender built from the same samples recovers the bit in 97.7\% of rounds, and every monitor without the true bit stays near chance against it. A canonical task interface preserves every receiver decision and bounds the remaining information about a balanced attribute by twice the routing-error rate, even for a receiver that shares a key with the sender. These results distinguish what passive auditing can certify from what restricting communication can enforce.

0.25

On Pearl Blockchain's Proof of Useful Work: Systematization and Flexibility

cs.CR cs.ET ·2026-10-18 ·Youlong Ding
unusual
Abstract

We revisit the proof-of-useful-work (PoUW) scheme in Pearl blockchain's white paper (Pearl Research Team), giving a systematic sensitivity characterization as an alternative to its Jackpot policy, along with efficient algorithm for verifying the sensitivity condition. Beyond its independent interest, our characterization yields a more flexible PoUW scheme that eenables miner autonomy: each miner can independently choose its noise magnitude to tailor the trade-off between computational accuracy and mining efficiency to its own needs.

0.25

Spectrum of Liability: Forensically Sound Acquisition of Wireless Evidence for the Reconstruction of Maritime Cyber-Physical Incidents

cs.CR ·2026-10-18 ·Giacomo Longo, Alessandro Cantelli-Forti, Alessio Merlo, Enrico Russo
Abstract

Modern maritime navigation depends on mandatory wireless inputs, most notably GNSS and AIS, whose reception can be spoofed or jammed. When groundings or collisions occur, liability and safety investigations often turn on whether RF interference corrupted bridge instruments, yet the artifacts typically available (VDR outputs, IBS logs, screenshots, decoded messages) are downstream, lossy, and easy to contest in adversarial proceedings. We introduce a forensically sound acquisition methodology that preserves upstream wireless evidence together with the minimal contemporaneous context needed to interpret it. Our system records coherent, multichannel IQ across relevant bands while simultaneously collecting time-aligned receiver and instrument-state metadata (e.g., decoded AIS/GNSS outputs and NMEA sentences), enabling later re-decoding and physical-layer reanalysis tied to what the onboard stack indicated at the time. To make the record independently auditable, we package samples and metadata into a single append-only forensic envelope that cryptographically binds per-stream hash chains with trusted-time anchors and periodic checkpoints, allowing third parties to detect tampering (modification, truncation, splicing, cross-case substitution) and to verify the claimed recording window. A prototype implementation shows the approach is practical: integrity mechanisms add negligible storage overhead and modest compute cost at operationally relevant multiband capture rates.

0.25

Equivalent Non-linear Shift Register Transformations and Espresso

cs.CR ·2026-10-18 ·Anirban Ghatak, Kaushik Sarkar, Anupam Chattopadhyay, Ambrish Awasthi, Indivar Gupta
unusual
Abstract

The stream cipher Espresso was designed by E. Dubrova and M. Hell as a hardware-efficient candidate for 5G security. It has two components - a 256-bit binary non-linear feedback shift register (NFSR), with 14 feedback taps, and a 20-variable non-linear output function. However, the security of Espresso with regard to several standard cryptanalytic attacks are estimated using an alternate NFSR, which has only two feedback taps. The above analysis is based on the claim that this alternate NFSR is obtained using a transformation algorithm which preserves the output sequence of the original NFSR. This issue has become all the more significant in the light of recent results on hardware performance evaluations, which assume that the stated equivalence holds. This article presents the first successful attempt in settling the above question of Espresso NFSR equivalence. We arrive at our main result by a careful analysis of existing transformation methods relevant to the question. In doing so, we first identify and correct a critical flaw in the proof of a foundational result of E. Dubrova. Next, in the context of Espresso equivalence, we analyze and unify the transformation framework of Dubrova and the one proposed by Yao and Parampalli. Based on our analysis, we have succeeded in transforming the LFSR in an Espresso-equivalent design reported by Yao and Parampalli to the Espresso alternate NFSR. Further, we have obtained a corresponding non-linear output function which achieves output-equivalence with the LFSR design, and hence, with Espresso.

0.25

PRICE: Privacy-Preserving Image Classification in Mobile Edge-Cloud Computing with Homomorphic Encryption

cs.CR ·2026-10-18 ·Duc-Huy Nguyen, Bui Duc Manh, Chi-Hieu Nguyen, Dinh Thai Hoang, Diep N. Nguyen +2 more
Abstract

Cloud-assisted image classification enables many mobile AI applications, but transmitting raw images to the cloud raises significant privacy concerns. Homomorphic encryption (HE) allows computation over encrypted data; however, existing end-to-end HE-based inference pipelines incur prohibitive latency and computational overhead, limiting their feasibility for latency-sensitive mobile systems. This paper addresses how to enable efficient encrypted image classification in mobile edge-cloud environments without sacrificing accuracy. We propose PRICE, a three-tier edge-cloud architecture that partitions deep neural network inference between a trusted edge and an untrusted cloud. The edge performs lightweight feature extraction in plaintext, while only compact embeddings are encrypted using the Cheon-Kim-Kim-Song (CKKS) scheme and offloaded to the cloud for encrypted final-layer inference. This design avoids homomorphic evaluation of deep backbone layers, significantly reducing computation depth, ciphertext size, and end-to-end latency. To further improve efficiency, we develop RowHyDiag, a row-packed hybrid diagonal encoding with hoisted-rotation-aware encrypted matrix-vector computation, which minimizes rotation operations and improves slot utilization. Extensive experiments on real devices across five datasets demonstrate that, under the evaluated 128-bit-security configurations, RowHyDiag achieves up to 31.8$\times$ faster cloud inference, 85.3$\times$ fewer rotations, and 83.0$\times$ lower setup payload than the conventional Halevi-Shoup diagonal method. RowHyDiag also outperforms the PEGASUS linear-transformation baseline, providing 1.1-4.5$\times$ faster cloud inference and 1.2-3.4$\times$ lower setup payload. The results show that PRICE enables deployment-ready, low-latency encrypted inference for mobile edge-cloud systems under a trusted-edge threat model.

0.23

TRACER: An RDC-Triggered Trajectory Recovery Framework for Cyberattack-Resilient Adaptive Cruise Control

cs.CR eess.SY ·2026-10-18 ·Yicheng Yang, Tianyu Liu, Omar Tahri, Haoli Zhao, Yongfeng Dong +1 more
Abstract

The increasing connectivity of adaptive cruise control (ACC) systems exposes vehicle-state information to cyberattacks, potentially causing irrational car-following behavior and unsafe longitudinal dynamics. Existing trajectory recovery methods generally lack an interpretable criterion for recovery activation and remain vulnerable to attack-induced distribution shifts. To address these challenges, this paper proposes TRACER, a rational driving constraint (RDC)-triggered trajectory recovery framework for cyberattack-resilient ACC systems. TRACER employs a rationality-aware triggering (RAT) mechanism that activates recovery when the observed driving state violates the calibrated RDC bounds. It also introduces cross-branch robust imitation learning (CRIL), which combines a clean branch for learning nominal driving behavior with a noise-enhanced branch for improving robustness to corrupted observations. Cross-branch fine-tuning further aligns the robust recovery policy with the nominal driving-state distribution. Experiments across six cyberattack scenarios show that TRACER achieves reliable recovery triggering and outperforms state-of-the-art recovery baselines in terms of accuracy, stability, and safety. State-space analysis further confirms that TRACER mitigates attack-induced distribution shifts and restores coupled spacing-velocity dynamics, highlighting the effectiveness of robust policy learning for cyberattack-resilient ACC.

0.22

Many Repairs, One Direction: Steering Code Models Toward Safer Cryptographic API Use

cs.CR cs.LGcs.SE ·2026-10-18 ·Tianhe Lu, Sakuna Harinda Jayasundara, Eric Spero, Ewan Tempero, Robert Biddle +1 more
Abstract

Cryptographic API misuse in large language model-generated code can compromise software security. While many existing approaches improve code security by modifying model parameters or inputs, how representation changes associated with different cryptographic API repairs can support a shared intervention for safer code generation remains underexplored. To investigate this, we construct 48,000 matched program pairs across four programming languages and ten cryptographic API misuse categories. Each pair changes a specified API decision while preserving the programming task. Subtracting the misuse program's internal representation from its secure counterpart's gives a repair vector. Across five models, comparisons of mean repair vectors reveal stronger organization by misuse category than by language. These category-specific patterns also support a shared intervention. We introduce CrypSteer, which extracts a shared direction by normalizing repair differences and selecting a principal component aligned with repairs across misuse categories. With model weights fixed, adding one shared direction per model to its internal activations during generation improves aggregate target security rates by 5.28--15.00 percentage points over unsteered generation across languages and misuse categories. These findings connect the shared structure of cryptographic repair representations to a reusable intervention for safer code generation.

0.20

SEAL: Towards Persistent Tool Poisoning Attack against LLM Agents via Prefix-Eliciting Trigger

cs.CR ·2026-10-18 ·Nanjun Zhou, Yiming Li, Yukun Chen, Ting Qiao, Enhao Huang +1 more
Abstract

Tool poisoning has emerged as a critical security threat to LLM agents, where adversaries embed malicious instructions in tool descriptions or external resources that enter the agent's context, allowing adversaries to manipulate agent's tool selection process. However, existing tool poisoning attacks primarily rely on an idealized assumption that the attacks task effect within a single interaction round. We show that their effectiveness degrades sharply in realistic multi-round agentic settings. We further reveal that this degradation stems from the fact that agents do not consistently surface attack-relevant signals in their reasoning as the planning context evolves, thereby weakening the attack's influence across rounds. To investigate whether tool poisoning remains a practical threat in such realistic settings, we propose SEAL, a two-stage method for persistent tool poisoning. Specifically, SEAL first performs Prefix-Eliciting Trigger Optimization (PETO) to optimize context-specific triggers that elicit attack-relevant reasoning prefixes, improving effectiveness within individual contexts. It then applies Local-to-Universal Trigger Aggregation (L2U) to combine these local triggers into a single universal trigger, improving generalization across different contexts and execution rounds. Extensive experiments across seven representative multi-round scenarios and different models demonstrate the attack effectiveness of SEAL, its transferability across target models, and its resistance to potential defenses.

0.16

Systems & Networking 24

Cryogenic Recharacterization of a Commercial 22 nm FD-SOI Foundry Standard-Cell Library

cs.AR quant-ph ·2026-10-18 ·Mohamed Gharib, Rami Rasheedi, Salma Abdelzaher, Henry Shi, Mike Hammer +4 more
Abstract · Submitted to the 2027 IEEE International Symposium on Circuits and Systems (ISCAS 2027)

Wiring and bandwidth bottlenecks between a quantum processor and room-temperature electronics can be reduced by placing digital control at the cryogenic stage, but commercial foundry standard-cell libraries are not specified for operation at 4 K. A cryogenic recharacterization flow is therefore developed for a commercial 22 nm FD-SOI foundry standard-cell library using commercially available cryogenic compact-device models, while retaining the original cell circuits, layouts, logic functions, and implementation semantics and remaining compatible with the commercial synthesis, place-and-route, and signoff methodology used at room temperature. Body-bias dependence is quantified over 38,640 matched slew/load points at zero, +/-0.9 V, and +/-1.8 V body bias. Relative to full bias, zero-bias operation increases the median rising and falling delays by 38.6x and 33.0x, while approximately 90% of transitions exceed the foundry-defined maximum-transition envelope. The 0.8 V-to-0.4 V level shifter is even more sensitive to body bias: its worst-case falling delay decreases from 30.6 ns at zero bias to 0.56 ns at +/-1.8 V. For the outbound 0.4 V-to-0.8 V crossing, driving the 0.8 V receiver without a level shifter dissipates 57.7 uW of static power per bit at full bias, against 0.196 nW when level-shifted. In post-route analysis of the unchanged extracted physical database of a fabricated 32-bit ALU, the register-to-register setup boundary increases from 1.3 MHz at zero bias to 37.9 MHz at +/-0.9 V and 132.9 MHz at +/-1.8 V.

0.51

panvram: GPU-resident random access to a human pangenome cohort

cs.DC cs.PFq-bio.GN ·2026-10-18 ·Yakiv Shavidze
code/data
Abstract · 7 pages. Code, harnesses and logs: https://github.com/yasha1971-coder/panvram (commit 8f8e9e1); evidence archive DOI 10.5281/zenodo.23449044

Repeatedly sampling short sequence windows from a human pangenome requires an access representation as well as compression. panvram stores each assembly as a reference-relative block archive and pools these archives on a GPU. For 558 haplotype assemblies, the default q4k representation occupies 16.642 GB of persistent pools including the decoded reference; a packed-reference, compact-directory representation occupies 11.634 GB. These quantities are not peak application memory. In the C1 experiment, default and fully compact forms delivered 743,271 and 736,762 windows/s; a separate public Colab run delivered 741,931 and 680,760 windows/s. All are batch API measurements and must not be pooled as repeated trials of an identical environment. Full-decoding checks compared reconstructed FASTA checksums with the manifest. On the CPU, the M6 shuffled-window workload at the same cohort cardinality reported 26 requests/s for AGC and 145,222 requests/s for refrel3 q4k. This comparison concerns one-thread requests of 4 KiB on the recorded warm-cache workload, not all genome-access patterns. AGC provides the smaller incremental archive: 5.40 MB/assembly versus 13.04 MB/assembly for refrel3 q16k, with different exclusions; the size pair is not the same refrel3 configuration as the q4k timing pair. We derive an expected-work expression for fixed independent granules and distinguish it from an empirical median-latency predictor. Cohort-scale counter and DRAM measurements motivate a locality hypothesis, but do not establish a causal per-miss latency or independently reproduce a model-failure verdict. The window-cost verdict at N=558 is published with its harness and raw timings.

0.45

Heuristic-Driven Resource Optimization for Decoupled Memory Channels on FPGAs

cs.AR cs.DC ·2026-10-18 ·Arnab A. Purkayastha, Shobhit Aggarwal, Jay Tharwani
published
Abstract · 6 pages. Accepted version of a paper published in the 2026 IEEE International Conference on Interdisciplinary Approaches in Technology and M…

Determining OpenCL channel depths for Decoupled Access-Execute (DAE) architectures in High-Level Synthesis (HLS) is critical for balancing performance and hardware efficiency. Conventional static analysis tools often over-provision these memory channels to ensure performance safety, leading to significant resource stranding. This paper presents an expansion of an LLVM-based static analysis framework, originally proposed for automated DAE generation, by introducing a hardware-aware refinement layer. We identify a counter-intuitive Performance Inversion Paradox linked to baseline memory stalls, where high-stall, grid-based kernels reach peak performance at minimal channel depths and suffer up to a 61% throughput degradation when depths are increased. We propose a Heuristic Refinement Layer that optimizes OpenCL channel depths based on this correlation between baseline stalls and performance saturation. Experimental results demonstrate that our expanded framework successfully identifies the Minimal Peak Performance point across various benchmarks. By avoiding the performance traps of standard over-provisioning, we achieve a 2.52x improvement in Resource Efficiency for an illustrative HotSpot application. This approach maintains peak execution speedup while ensuring the leanest possible hardware profile, outperforming conservative tool-based static predictions.

0.40

Triple Overlap in Pipelined Krylov Methods on Multi-GPU Systems

cs.DC ·2026-10-18 ·Sinan Ekmekçibaşı, James D. Trotter, Johannes Langguth, Xing Cai, Didem Unat
publishedcode/data
Abstract · 9 pages, 6 figure. Includes artifact description and artifact evaluation appendices. Accepted to SC26-W, IA^3 Workshop

We introduce a novel triple-overlap implementation of the pipelined preconditioned conjugate gradient (Pipe-PCG) and pipelined stabilized biconjugate gradient (Pipe-BiCGStab) solvers, methods widely used in scientific applications. By leveraging a collective communication library, our approach extends conventional computation-communication overlap to concurrently execute collective operations, irregular P2P communication, and compute kernels, enabling computation-communication-communication overlap across multiple GPUs. We evaluate the design on the MareNostrum 5 (H100) and Leonardo (A100) supercomputers, comparing against PETSc and the state-of-the-art two-stream NCCL implementation using SuiteSparse matrices and three massive instances scaled up to 512 GPUs. The three-stream design outperforms the two-stream baseline on every instance, with the margin widening from 2-6% at 4 GPUs to 18-47% at 64 GPUs. At 64 GPUs, the geometric-mean speedup over the two-stream baseline is 1.33x on MareNostrum 5 and 1.23x on Leonardo for Pipe-PCG, and reaches 1.30x for Pipe-BiCGStab. In large-scale experiments, we observed a speedup up to 35% in Pipe-PCG and up to 33% in Pipe-BiCGStab compared to the two-stream design. For PETSc, the margin is even wider, up to 3.27x speedup in both Pipe-PCG and Pipe-BiCGStab. Related code for the proposed method is available online: https://github.com/ParCoreLab/aCG

0.40

Which Code Changes Need Re-embedding? CIPBench and the Limits of Multi-Signal Index Maintenance

cs.DB cs.SE ·2026-10-18 ·Nannan Zhao, Jiameng Zhang, Taoyu Zhong, Hao Wang, Bo Wang
benchmarkcode/data
Abstract · 39 pages

Repository-level retrieval-augmented generation systems must keep vector indexes synchronized with continuously evolving code, yet there is no principled basis for deciding which changed code units actually need new embeddings. We ask whether the retrieval impact of a code change can be anticipated before that embedding is computed. Across 10 open-source repositories in Python, Go, and JavaScript, we characterize how a change propagates from syntactic substantiality through embedding displacement to retrieval impact. The links are informative but not deterministic: syntactic class separates displacement sharply, yet impact depends on a unit's retrieval exposure as much as on how far its representation moves, and recurring counterexamples bound what any single signal can decide. We formalize the decision as CIPBench, a benchmark that scores maintenance policies under a strict pre-embedding information boundary. Prediction works, and one signal exhausts it. A ranker using only a unit's change history substantially reduces retrieval harm relative to random and recency-based scheduling. Combining it with three further signals--individually discriminative and largely unassociated with one another--leaves the decision statistically unchanged across every budget we evaluate, for reasons we trace to the structure of the score rather than to its tuning. What pays off is not ranking but filtering: units whose text is unchanged from the indexed state can be skipped with no risk and, in the most reversion-heavy project, at nearly half the embedding cost. Signal-level informativeness is therefore not evidence of decision-level value, and separating the two requires evaluating the decision a signal produces rather than its association with the outcome.

0.37

gKSM: GPU Memory Deduplication at Driver Layer for Serverless LLM Inference

cs.OS ·2026-10-18 ·Shuheng Gao, Fan Guo, Yongkun Li, Ziye Yue, Yinlong Xu +1 more
published
Abstract · Accepted by SIGOPS ATC 2026

In serverless LLM inference, cloud providers commonly locate multiple instances on a single GPU to improve utilization. However, each instance independently maintains its model weights and runtime state, making duplicate GPU memory the primary bottleneck. Memory deduplication is a widely-used technique for eliminating duplicate data in CPU systems, but extending it to GPU memory faces fundamental challenges. Effective GPU memory deduplication must avoid redundant data movement between CPU and GPU while preserving write isolation for shared pages without native copy-on-write (COW) support. This paper presents gKSM, a transparent GPU memory deduplication framework for serverless LLM inference. gKSM addresses these challenges through CPU-GPU collaborative hash management that offloads hash computation to GPU kernels, zero-copy model weight deduplication that eliminates duplication before model weight transfer, and a cross-privilege page fault handler that combines a lightweight kernel hook and CUDA Virtual Memory Management (VMM) remapping in the user space. gKSM also designs a stable page scanning algorithm and a reuse-aware placement policy in cluster scale. Experiments show that gKSM reduces GPU memory usage by 59.8% while maintaining high throughput compared to ServerlessLLM. Under memory constrained workloads, gKSM improves aggregate throughput by 2.15X and supports five instances concurrently. In cluster level simulation, gKSM achieves a service rate of 99.97%, compared to 78.2% for SLInfer and 26.9% for ServerlessLLM.

0.35

Mixed-Precision GPU Acceleration for Confidence Regions of Spatial Excursion Sets

cs.DC ·2026-10-18 ·Xiran Zhang, Sameh Abdulah, Qinglei Cao, Joshua French, Rabab Alomairy +2 more
Abstract · 12 pages

A spatial excursion set identifies where a field exceeds a threshold. With incomplete and noisy observations, inner and outer confidence regions quantify uncertainty in its spatial extent through lower and upper bounds at a specified confidence level. Constructing these regions requires repeated conditional simulations accounting for joint spatial uncertainty, making dense covariance factorization and simulation major computational bottlenecks. We present a GPU-accelerated framework that combines adaptive tile-based mixed-precision arithmetic with distributed task execution to accelerate universal kriging, prediction-uncertainty estimation, conditional Gaussian simulation, and confidence region calibration. In our parallel mixed-precision simulation workflow, triangular matrix-matrix multiplication (TRMM) inherits the adaptive precision layout of the Cholesky factor and uses FP16 simulation operands, while covariance scaling and compensating operand rescaling accommodate the numerical range of reduced precision arithmetic. With appropriate precision tolerances, empirical coverage is close to the double-precision reference and the nominal 90% confidence level across weak, medium, and strong spatial correlation. Ocean surface-salinity experiments show close agreement with the double-precision reference in spatial predictions and confidence regions. On four NVIDIA A100 GPUs (40 GB each), mixed-precision TRMM reaches 6.36x to 6.93x the theoretical FP64 peak throughput. On four NVIDIA GH200 Hopper GPUs, it reaches 10.20x to 11.03x. Each AMD MI250X node has eight GPU dies. Aggregate throughput rises by 9.1x to 11.5x from 1 to 16 nodes under weak scaling and by 1.6x to 1.9x from 4 to 16 nodes under strong scaling. Adaptive mixed-precision accelerates large-scale excursion set analysis while preserving the statistical accuracy required for spatial uncertainty quantification.

0.34

An Information-Geometric Framework for OCS-Reconfigurable Interconnect Topology in Large-Scale AI Clusters

cs.DC cs.LG ·2026-10-18 ·Shintaro Yoshizawa
theoryunusual
Abstract · 27 pages

Optical Circuit Switches (OCS) make inter-pod wiring in AI clusters programmable, posing design questions on wiring choice, symmetry departure, and reconfigurations. We present a unified information-geometric framework to address these issues.A wiring is a point in a probability simplex over circuits within a degree polytope. Departure from a symmetric Fat-Tree prior $\mathbf{q}$ is measured by $\DKL(\cdot \,\|\, \mathbf{q})$. The cost $U$ is the smoothed DAG critical-path length, whose gradient equals path marginals of a Gibbs distribution via forward--backward recursion.For free energy $F_{β_1} = U + β_1^{-1} \DKL(\cdot \,\|\, \mathbf{q})$, we prove well-posedness, monotonic dependence on $β_1$, adaptation bounds via Jeffreys divergence, and a symmetry principle for static optimality. An entropic proximal-gradient method updates multiplicatively with a $|V|$-dim dual KL projection, achieving linear convergence.For time-varying traffic, a threshold policy yields regret $O(\sqrt{ΓC V_1 T})$ against the switching-free optimum. Traffic changes are detected by a multinomial GLR--CUSUM detector with non-asymptotic false-alarm bounds. Multi-job generalized-mean fairness aligns with $α$-fairness ($α\ge 2$). Experiments on pipeline, data, and MoE workloads validate the framework against integer programming.

0.31

Ineffective Occupancy in Heterogeneous GPU Clusters: Prediction, Contribution Analysis, and Readiness-Aware Scheduling

cs.DC ·2026-10-18 ·Tao Zhang, Bin Liao
theorycode/data
Abstract · 24 pages, 10 figures, 13 tables. Code: https://github.com/liaobin665/gpu-ras

In heterogeneous GPU clusters, GPUs are costly and scarce, yet placing a job on a machine does not mean that the job becomes ready to run. A large amount of allocated capacity can remain occupied for long periods without producing useful computation, further aggravating the waste of compute resources. On the Alibaba GPU cluster trace we find that, even when GPU capacity is sufficient and placement is complete, 52.70% of execution records never became ready. We term this phenomenon ineffective occupancy. Building on this diagnosis, and keeping the placement policy fixed, we propose Readiness-Aware Scheduling (RAS): it uses only nine attributes observable at admission time to predict readiness probability and reorders admission accordingly. Under a five-fold split grouped by machine, the gradient-boosted model reaches an ROC-AUC of 0.8187. A feature-group contribution analysis shows that the model's score is driven mainly by job profile, duration, and GPU model, with the request-size group contributing least, followed by the scheduling and concurrency group. In GPU terms, RAS reduces the ineffective-occupancy ratio (the share of cluster GPUs held under ineffective occupancy) from 43.64% to 27.15% and raises the effective coverage of large requests (eight GPUs or more) from 46.11% to 63.18%, without sacrificing capacity utilization. Relative to the upper bound obtained by ordering with ground-truth labels, RAS captures 65.1% of the attainable improvement, well above the interpretable rules built only from the contribution-analysis conclusions; the gain therefore comes mainly from predicting with all nine attributes jointly rather than from combining rules built on any single attribute group. These results indicate that readiness status is a measurable, predictable, and optimizable dimension of admission decisions in heterogeneous GPU clusters.

0.29

Ride, Raise, or Wake: A Theory of Light-Thread Placement on Multi-Tier Processors

cs.PF ·2026-10-18 ·Mohamed Amine Bergach
Abstract

hones and laptops now mix several kinds of cores, and each kind sits in its own voltage and clock domain. When a light thread arrives, should it run beside heavy threads on a fast cluster that is already awake, or wake a cluster of efficient cores? We answer with three costs. A thread can \emph{ride} an awake domain at the voltage other threads hold, \emph{raise} that voltage for everyone, or \emph{wake} a sleeping domain and pay its fixed power. From these costs follows a simple rule: ride while the thread's demand is below $r^*=Φ/(e_h-e_l)$, the fixed power of the cluster it would wake over the gap in energy per operation; otherwise wake the smallest tier that can serve it. The best tier is not monotone in demand: tiny threads ride high, medium threads wake a low tier, heavy threads go high again. Two common designs err on opposite sides of the rule: grouping a light thread with the heavy threads of its application overpays above the threshold, and an energy model without idle costs overpays below it. In a model with five tiers, grouping costs 80 to 145\% more energy than the rule, and more as tiers are added. Measurements on an Apple M6 anchor the parameters.

0.26

StreamAhead: Real-Time Streaming Video Question Answering via Adaptive Precomputation

cs.DC cs.CV ·2026-10-18 ·Haoyang Fan, Dhruv Parikh, Rajgopal Kannan, Viktor Prasanna
Abstract · 12 pages, 5 figures, 6 tables

Real-time streaming video question answering with vision-language models must answer queries that arrive at any time with low latency while sustaining the video-processing rate required by the application. Existing systems construct key-value (KV) caches for the entire video as it arrives, although a query typically uses only a few of the preceding frames as evidence. This highlights a fundamental trade-off: performing more computation as video arrives reduces query-time computation but can increase processing lag, whereas deferring computation can increase the video-processing rate but requires additional query-time computation for retrieved content. We present StreamAhead, an algorithm-system co-design that exploits this trade-off through adaptive precomputation: dynamically controlling how much computation to complete for each video segment before a query arrives. StreamAhead encodes video segments independently and lets each remain at a different representation level, from inexpensive visual features to costly KV caches, completing KV caches only for the retrieved segments when a query arrives. StreamAhead profiles offline the cost of advancing a segment to each representation level. Its runtime uses these profiles to schedule segment computation concurrently with query execution, pipelining it across heterogeneous CPU and GPU resources while maintaining the required video-processing rate. StreamAhead achieves accuracy competitive with state-of-the-art methods with up to 10.4x lower time-to-first-token (TTFT), and up to 5.9x lower TTFT than state-of-the-art streaming systems. It sustains 30 FPS with P95 TTFT under 0.5 s, and up to 57 FPS with P95 TTFT under 2 s. To our knowledge, StreamAhead is the first real-time streaming video QA system with competitive accuracy, setting a new Pareto frontier between video-processing rate and TTFT.

0.25

RemoteANN: A Native Vector Search Engine for Cloud Object Storage

cs.DB cs.IR ·2026-10-18 ·Yutong Gou, Silu Huang, Yue Wang, Jianjun Chen, Cheng Long
hot direction
Abstract

Vector search engines built on cloud object storage have drawn growing attention from both academia and industry. These systems typically adopt a disaggregated storage-compute architecture, persisting the primary vector payload in object stores (e.g., Amazon S3) while retaining compact index metadata on compute nodes. However, conventional vector indexing structures are not natively tailored to the pronounced network latency and per-request access costs of cloud object stores, resulting in network I/O amplification, degraded query throughput, and higher API expenses. To resolve these challenges, we present RemoteANN, a vector search engine explicitly co-designed for cloud object storage across indexing and querying. To mitigate the dominant network I/O bottleneck, RemoteANN couples an Inverted File (IVF) index with RaBitQ quantization, further optimized by inter-cluster and intra-cluster data reordering for improved data locality. Building on this tailored data layout, RemoteANN's query engine combines cluster pruning, fine-grained Range-GET requests, request merging, and local NVMe SSD caching, dramatically reducing discrete storage requests and network transfer volume without sacrificing accuracy. To address the index construction bottleneck, we further develop QGKMeans, which substantially reduces construction latency without compromising clustering quality. Extensive evaluations on real-world datasets demonstrate that, compared with representative baselines, RemoteANN achieves up to a $5.9\times$ throughput speedup while issuing up to $16.1\times$ fewer object storage requests at 97% target recall.

0.25

BitPCE: Batched Multi-Dimensional PolicyResolution with GPU Bitmap Algebra

cs.DC cs.CR ·2026-10-18 ·Shengwei Wang
Abstract

Policy controllers must determine which workloads match rules spanning several label dimensions. Hash sets, compressed bitmaps, and GPU set-processing primitives address the underlying set operations. However, intermediate results, repeated scope evaluations, dependent policy updates, and host transfers can dominate the cost of a heterogeneous batch. We study this problem as batched intersection-of-unions (IoU). Our design exploits independent scope--word tasks, reusable endpoint expressions, and policy dependencies. The challenges are preserving wildcard, empty-group, and exclusion semantics, recognizing equal scopes, and retaining unchanged recipients affected by source edits. BitPCE addresses them with register-resident GPU evaluation, canonical scope compilation, and dependency-complete candidate selection under a stated rendering contract. IP-only changes reuse cached memberships. We establish correctness and word-work bounds. On a four-vCPU A10G host, resident GPU execution is 3.46--36.60$\times$ faster than the best tested native CPU path, including CRoaring, and 1.14--2.22$\times$ faster than an adapted operand-parallel GPU reduction. Gains narrow at the host-output boundary: on an eight-vCPU host, including upload and compact identifier output, resolution takes 16.3 ms versus 19.3 ms for CPU at 100K objects and 500 scopes. Sparse labels and small batches can favor CPU execution. Scope reuse, candidate selection, and incremental transfers are measured as components; complete update-to-enforcement latency remains unevaluated.

0.21

Queue-Aware ISAC Power Allocation Under Imperfect CSI: Integer-Service Analysis and Anchored Lyapunov Look-Ahead

cs.NI ·2026-10-18 ·Hanyi Sun, Yifan Zhou, Yixiao Gu
Abstract

We study power allocation for a multiuser integrated sensing and communication (ISAC) downlink with bursty queues and imperfect channel state information (CSI). Starting from a long-run utility--backlog objective, we derive its common task reward and online score. Integer departures are represented by threshold events of indefinite Gaussian quadratic forms, accounting for estimation error in both signal and interference. Their moments give the conditional one-slot task reward and characterize the error of an efficient plug-in score. An anchored quadratic Lyapunov look-ahead corrects its reward-greedy action with a bounded nominal loss. Across five CSI-error levels, the rule reduces mean queue exceedance by 27.4--43.9\% (0.49--0.99 percentage points) relative to the plug-in maximizer, with mean reward loss below 0.01 per slot. The rule achieves higher reward and lower exceedance than DPP, SAC and PPO; its advantage over freshly retrained learning baselines extends to five configurations with 9--23 power-allocation variables using fixed candidate budgets. For the seven-variable system, a fused implementation reproduces all 25 confirmation trajectories and achieves a one-core p99 decision time of 0.274--0.311\,ms.

0.21

MemDerive: A Diagnostic Benchmark for Agent Memory over Evolving Data

cs.DB cs.AIcs.CLcs.IR ·2026-10-18 ·Sen Gao, Yongan Li, Han Chen, Bingsheng He, Ming Wu +3 more
benchmark
Abstract

Long-running agents rely on memory of their past interactions for task state, such as what has been completed and which conditions still hold. Such state is usually a derived state. No record states it, and it must be computed by rules from multiple records of an evolving history. Existing memory benchmarks score only the final answer, with no reference for the intermediate states behind it. A wrong answer therefore cannot be traced to the state that was missed, and a state that silently became outdated goes undetected. We present MemDerive, a diagnostic benchmark for agent memory over evolving data. It specifies each scenario symbolically, derives exact gold answers with Datalog, records the gold derivation DAG and the supporting events of every query, and renders the history in natural language with fictional entities and distractors. Its 10 scenarios contain 4,692 events and 1,503 queries whose difficulty varies along access, evolution, and composition. On this ground truth, MemDerive evaluates answer reliability, diagnoses failures by probing the nodes of the gold derivation DAG, and tracks whether answers stay correct as events arrive. Our evaluation of four memory systems reveals failures that final-answer benchmarks hide. On derived state, every memory system falls behind simple BM25 retrieval, although the same answer model reaches 96.01% when given the supporting events. Most failures are ungathered, as a system holds the missing intermediate state but does not retrieve it for the question that needs it. Memory systems also rarely propagate implicit changes, which no new record states. These results suggest three principles for memory design. Gather inputs along the derivation, store rules with their versions, and maintain derived states through recorded dependencies.

0.21

Off-Tree Forwarding for Accelerating Broadcasts from Arbitrary Sources in Plumtree

cs.DC cs.NI ·2026-10-18 ·Ryohei Banno
Abstract

Overlay broadcast is fundamental to large-scale distributed systems, enabling messages to be disseminated to all participants. Plumtree is a well-established protocol that embeds a spanning tree in a gossip overlay: payloads are disseminated along tree links, while non-tree links support recovery and optimization. However, because the tree is constructed from the first broadcast, it is optimized for the initial source and can introduce long detours for broadcasts originating at other nodes, particularly leaf nodes. To address this limitation, we propose probabilistic exuding, an off-tree forwarding mechanism in which nodes near the source additionally forward payloads over their non-tree links with a specified probability. Exuded messages use a dedicated message type and do not participate in tree construction or repair, leaving the shared tree unchanged. Simulations on a 1000-node random regular overlay show that, for randomly selected sources and a representative parameter setting, the proposed method reduces the 95th-percentile delivery latency by approximately 40% relative to Plumtree. This closes approximately 92% of the latency gap between Plumtree and flooding while sending only about 58% as many payload messages as flooding. In contrast, Plumtree's own link optimization and shorter repair timeouts, which also exploit non-tree links, rewire the shared tree and increase latency as broadcast sources change; the latency of exuding is nearly unaffected.

0.21

VS-Flash: Harnessing Vector Search with Computational Memory-Semantic Flash

cs.DB cs.DC ·2026-10-18 ·Donguk Kim, Sangjun Pae, Daeyoung Lee, Dong Kim
Abstract

As generative AI continues to evolve, retrieval-augmented generation (RAG) systems are driving rapid growth in vector database capacity requirements. As datasets outgrow host DRAM, vector databases rely on MMap-based serving over SSDs to extend capacity. However, high-fan-out ANN searches generate concurrent irregular accesses that contend for host memory and storage resources, degrading bulk throughput and inflating tail latency under mixed workloads. In this paper, we address these challenges through memory-semantic capacity and workload-aware compute placement. VS-Flash exposes flash as a large-capacity memory-semantic space and dynamically selects the execution engine according to runtime query and workload characteristics. It offloads throughput-oriented large-batch queries to a Computational Memory-Semantic Flash (CMS-Flash) device, while executing latency-sensitive small-batch queries on the host CPU. Both paths access the same resident segments, avoiding a separate host-DRAM replica. Across three billion-scale datasets, VS-Flash improves bulk-search throughput over vanilla Milvus by 36.8X on average, with up to a 57.4X improvement. Under mixed workloads, VS-Flash reduces the p99 latency of latency-sensitive requests by up to 24.8X while sustaining device-side bulk search. These results show that VS-Flash delivers scalable vector search beyond host DRAM capacity while jointly improving bulk throughput and small-request tail latency.

0.20

Visual Cache Reuse for Efficient Multimodal Inference

cs.OS ·2026-10-18 ·Shuheng Gao, Yuankai Fan, Jiaqiang Ma, Jiawei Shao, Yongkun Li +1 more
Abstract

Vision-language model (VLM) inference incurs substantial prefill latency and memory overhead due to the long visual token sequences of large visual inputs. Existing serving systems typically exploit exact cache reuse through prefix caching or position-independent caching, overlooking the potential of approximate caching in real-world workloads. However, directly reusing an approximate cache entry may introduce spurious visual information, which naive approaches that select a subset of important visual tokens for recomputation fail to adequately mitigate this risk. We present VisMemo, an efficient serving system that enables approximate KV cache reuse for modern VLMs. Its core is a system-algorithm co-design that combines an entropy-aware KV cache reuse algorithm with differentiated cache management. Specifically, VisMemo uses patch entropy to identify reuse candidates and warmup attention to prioritize recomputation based on the current query. To support the large cache capacity across GPU memory, CPU memory, and SSD storage, VisMemo applies distinct I/O and cache policies to exact and approximate hits: sparse patch-level I/O for approximate hits, prioritized retention for exact hits, and a reserved CPU victim cache for preset entries selected using offline workload priors and online reuse observations. VisMemo further optimizes the memory access path between the inference engine and the storage backend, eliminating redundant caching and unnecessary data movement. Experiments on multiple public datasets show that VisMemo preserves output quality under approximate reuse. On a real-world production trace, VisMemo increases the overall cache hit rate from 27.78% to 61.45% and reduces prefill latency by up to 48.3% at 8K input tokens.

0.20

Data.Readme: Documenting Table Relationships for Diverse Data Tasks

cs.DB ·2026-10-18 ·Jiaming Liang, Yeye He, Haoyu Dong, Zachary Ives, Surajit Chaudhuri
Abstract

Large language models (LLMs) can perform diverse data tasks when given a small set of relevant tables. However, for complex tasks requiring multiple tables over large data lakes, identifying relevant tables for the task becomes a prerequisite, but remains challenging. This is because there are often many tables in a data lake with similar table and schema names, and limited additional semantic context, thereby forcing agents to spend substantial time exploring tables at online query time, greatly affecting models' ability to perform tasks accurately, while increasing cost/latency substantially. We present Data.Readme, a novel table documentation approach that uses lightweight offline processing to infer and index rich \emph{table-to-table relationships}, including confusable, joinable, and union-able tables. We further develop a suite of agentic tools for LLMs to work with the index, to help agents retrieve relevant tables and perform tasks much more efficiently and effectively. Finally, we study how to optimize our index construction with a budget constraint. On data lakes containing up to 12K tables, Data.Readme improves task accuracy by as much as 40 percentage points across a range of data discovery, data analysis, and data science tasks, outperforming diverse baselines, including powerful coding agents with unlimited time and monetary budgets. We also show that equipping existing methods (such as coding agents) with Data.Readme can not only lead to consistent quality gains, but also significantly reduce their latency by up to 6x, demonstrating the potential of Data.Readme.

0.19

Uber-Join: Unifying Diverse Table Joins using LLMs with Precision and Cost Guarantees

cs.DB ·2026-10-18 ·Guohui Guan, Chang Ge, Yeye He
benchmark
Abstract

When two tables from different sources are joined, the values that should join often differ in spelling, format, or meaning, and fuzzy-, transformation-, and semantic-joins each handle one type of these differences. Large language models (LLMs) can match values that differ in several ways at once, but every LLM call is billed, and having an LLM compare the rows of two large tables can cost far more than a typical budget allows. This paper studies how to combine these join types with LLMs to join two tables under a dollar budget, maximizing recall while keeping the precision of the result at or above a target, a problem that is NP-hard. Our key idea is to use LLMs selectively to create join programs and verify their outputs through sampling, while adaptively deciding which joins to pursue under budget and precision constraints. This combines the scalability of program-based joins with selective LLM verification, while maintaining a statistical precision guarantee on the accumulated result. We develop Uber-Join, which treats creating a join and checking the output of a created join as two levels of arms in a budgeted multi-armed bandit. It commits an output only if the precision of the whole result, estimated from an LLM's judgments of sampled pairs, stays at or above the target with high confidence. Under the same budget, Uber-Join achieves the highest recall at the precision target in nearly all settings on three join benchmarks and Join-Hard, a new benchmark of large tables.

0.17

From Last-Layer to Cross-Layer: One-Shot Joint Classifier Reconstruction for Federated Learning

cs.DC ·2026-10-18 ·Zhenyan Liu, Hua Zhang, Haoran Gao, Huiyu Zhou, Zongliang Shen +4 more
Abstract · 9 pages, 5 figures

Federated learning (FL) enables collaborative model training while preserving data privacy. However, non-independent and identically distributed (non-IID) data inevitably introduces classifier bias, leading to significant performance degradation. Existing studies primarily focus on calibrating or reconstructing the classifier after FL training to alleviate this bias. We find that the last-layer representations of the global model are often suboptimal for establishing a global decision boundary. Instead, classifiers constructed based on intermediate-layer representations can achieve superior performance. This finding provides a critical insight into overcoming the performance bottleneck inherent in existing paradigms. Nevertheless, identifying the optimal layer without auxiliary datasets or prior knowledge remains a formidable challenge. In this paper, we propose FLARE (Federated Cross-Layer Joint ClAssifier REconstruction) which rethinks classifier reconstruction from a cross-layer perspective and generalizes ridge regression to jointly exploit cross-layer representations, thereby surpassing the performance of any individual layer. Specifically, after FL training, clients perform only a single forward pass to extract class-conditional statistics from predefined layers. The server aggregates these statistics to construct a joint cross-layer ridge classifier, where an adaptive coupling coefficient regulates cross-layer correlations, thereby facilitating the integration of complementary discriminative information. Extensive experiments across 11 datasets, 8 model architectures, and diverse data heterogeneity scenarios demonstrate that FLARE, as a plug-and-play post-processing method, improves accuracy by 11.35% with only one additional communication round.

0.17

Hierarchical Rule Auditing with Exact Ternary GPU Correlation

cs.DC cs.CR ·2026-10-18 ·shengwei wang
theory
Abstract

Auditing firewall logs against requirements mapped to controls and frameworks must retain overlapping obligations, preserve uncertainty from missing attributes and count distinct affected records. Direct, bitmap and relational methods can express this audit; a correlation implementation must preserve distinctions that ordinary match counts and rule-hit totals discard. We present an exact ternary correlation algorithm that compiles categorical conjunctions into one bounded-integer matrix product. A dominating false-predicate penalty separates known mismatches from missing evidence, while per-record hierarchy unions prevent duplicate counting. We prove their composition and handle high-cardinality endpoints through exact address classes and versioned labels. Evaluation compares eight CPU/GPU implementations, includes public traffic and 133 published source-address ACLs with generated observations, and measures hierarchy and policy-rebuild costs. On an NVIDIA L4, matched fused execution reduces latency by 8.8 to 11.2 percent relative to two-product correlation across five tested shapes. At 100,000 records and 1,024 requirements, it takes 12.28 ms versus 13.82 ms for two products and 12.29 ms for GPU bitmaps. Bitmaps remain faster in most regimes, with hierarchy operations and reductions dominating the measured fused-correlation stage intervals.

0.16

PlaneZip: Lossless Bit-Plane Compression of Training Activations Offloaded over PCIe

cs.DC ·2026-10-18 ·Zhufan Gu, Qizhe Tang, Yuxin Wang, Xun Zhou, Hui Feng
Abstract

Training under a device-memory budget relies on tensor rematerialization: an evicted activation is either swapped to host memory or dropped and recomputed. Because GPU compute has outgrown PCIe bandwidth, cost-model planners such as MegTaiChi, which charge a swap its full transfer time, almost never swap, and swaps that fill idle link time carry only a small fraction of the saved activations. Saved activations, however, are homogeneously typed arrays with bit-level redundancy beyond zeros, which lightweight bit-plane compressors can remove at line rate. We present PlaneZip, which compresses swapped tensors on the copy path and plans around the compression. Its codec picks, for every 128-byte block, the smallest of zero-value, delta bit-plane, bit-plane, and a new zero-masked bit-plane encoding. A per-tensor predictor reuses the ratio of the previous iteration, and the planner budgets link time in compressed transfer times. On activations captured from five models trained on real data, the codec compresses BF16 activations by 1.71x and FP32 activations by 1.39x (geometric means), against 1.18x and 1.20x for zero-value compression. In trace-driven simulation of a PCIe 3.0 consumer GPU and a PCIe 4.0 datacenter GPU at half the activation memory, PlaneZip cuts the rematerialization overhead by 47% and 31% relative to a bandwidth-aware planner without compression, and by 73% and 54% relative to MegTaiChi's cost model, with an estimated engine area of 0.036 square millimeters.

0.15

AgentLens: Framework-Aware Characterization and Resource Control for Heterogeneous LLM Agents

cs.OS ·2026-10-18 ·Wentao Hong
Abstract

LLM-based agents are increasingly deployed on shared hosts, yet their host-side resource demands remain poorly understood, limiting effective resource coordination. We present AgentLens, which combines a cross-framework workload characterization with a framework-aware resource controller for co-located agents. Across 10 frameworks and 11 matched coding tasks, we find that framework orchestration strongly shapes resource demand: framework identity explains 52.95% of peak RSS variance, compared with 4.22% for task identity, while execve events per tool call differ by 39x between MetaGPT and CAMEL. Resource dimensions are not interchangeable: peak RSS and subprocess count have a Spearman correlation of 0.13. Static budgets and framework-agnostic policies miss these structural differences and the alternation between API waiting and tool execution. AgentLens combines offline framework fingerprints with online execution-phase signals to coordinate memory through cgroup v2, without modifying framework internals. In trace-driven replay across three co-location pairs at a memory budget of 90% of aggregate peak demand, AgentLens keeps the protected worker's P95 allocation latency at 5.7-7.1 ms, improving on static allocation by 2.1-2.6x.

0.07

Software & Programming 12

Theory of Developer Momentum: The Developer's Experience of Progress and the Craft of Maintaining Motion

cs.SE ·2026-10-18 ·Arty Starr, Keith Mann, Margaret-Anne Storey
Abstract · 44 pages + 19 pages of appendix, 8 figures, 2 tables

This paper introduces a Theory of Developer Momentum that helps us see the activity of developing software in motion and the skilled craft through which progress unfolds. We define Developer Momentum as the developer's experience of progress and forward motion toward the goal of a task. Our theory disentangles the related constructs of flow state, momentum, and motivation, so we can more clearly see the relationships between them and how they interweave to support progress. In order to make progress, developers build up working context in memory, become immersed in a cognitive puzzle, and gain clarity through motion in uncertain terrain. The theory provides researchers with a rich experiential lens to investigate how developers' cognitive experiences are shifting in emerging AI workflows. Building on prior grounded theory work, we conducted in-depth interviews with a theoretically sampled subset of 12 professional developers, and eight feedback interviews to evaluate resonance and investigate how the theory translates to agentic AI workflows. We followed a Constructivist Grounded Theory approach to construct a theory grounded in empirical data. Our theory contributes a novel paradigm for seeing the motion of developing software at multiple scales and resolutions, seeing the living capacities that support this motion, and expanding what becomes available for improvement, skilled judgment, and craft.

0.40

Beyond Similarity-Based Generation: Guiding Commit Message Generation with Issue-Commit Traceability

cs.SE ·2026-10-18 ·Chunying Zhou, Gong Chen, Xiaoyuan Xie, Yi Song, Yuxuan Yang +1 more
code/data
Abstract · 16 pages, 5 figures

Commit messages provide records of code changes and help developers understand and maintain software. To reduce the manual effort involved in writing them, commit message generation (CMG) approaches automatically generate commit messages that describe code changes in natural language. Existing CMG approaches generally treat CMG as a standalone task and optimize generated messages primarily for similarity to developer-written references. However, this objective may not properly reflect the practical usefulness of generated commit messages. This limitation matters in current AI-assisted software engineering workflows, where commit messages are used as shared information to serve different software engineering activities. To fill this gap, we propose TraceMsg, which uses issue-commit traceability, a representative software engineering activity, to guide commit message generation. TraceMsg is a collaborative multi-agent framework that coordinates message generation, traceability reasoning, and feedback-guided refinement, allowing generated messages to serve traceability while using reasoning feedback to guide issue-conditioned refinement. We evaluate TraceMsg on six open-source projects. The experimental results show that TraceMsg achieves an average Hit@1 of 0.893 and an MRR of 0.911, improving Hit@1 and MRR by 13.16% and 38.75%, respectively, over the best baseline for each metric.

0.35

Neutrosophic Graph Rewriting as a Computational Model

cs.LO ·2026-10-18 ·Apostolos Syropoulos, Basil Papadopoulos, Binod Kumar Sharma
publishedtheoryunusual
Abstract

Graphs, which are made up of points connected by lines, are used to represent relationships between objects. They are useful for solving problems in computer science, logistics, biology, and other fields, including social networks and the internet. Neutrosophic graphs are graphs whose points and lines are characterized by membership, indeterminacy, and non-membership degrees. In the simplest case, only lines are characterized by these degrees. The process by which we transform one graph to another is called graph rewriting. It has been shown that graph rewriting can be described mathematically by pushouts, which are category-theoretic structures. It has also been shown that fuzzy graphs have this property, and here we prove that neutrosophic graph rewriting can be described by pushouts in a certain category. By providing a fuzzy version of category theory, it is possible to define fuzzy categorical structures and thus to define neutrosophic graph rewriting in a more suitable manner. Since graph rewriting is actually a model of computation, neutrosophic graphs are a neutrosophic model of computation. A hypergraph is a generalization of a conventional graph. It allows a single "hyperline" to connect many points, which is different from regular graph lines that only connect two points. Neutrosophic hypergraphs can be used to reformulate Wolfram's models, which are supposed to be a mathematical model of reality.

0.32

A sound pure type system with ranks as sorts and Type-in-Type

cs.LO math.LO ·2026-10-18 ·Pavel Sokolov, Alexander Kuklev
code/data
Abstract · 28 pages, 3 figures, supplementary material to be found at http://github.com/TurtlePU/pts-rk

We present a dependently typed pure type system that contains a circular axiom $s : s$ for a certain sort $s$ and is nonetheless strongly normalizing. Girard's paradox is avoided because our sorts are not closed under a $Π$-former: function types are placed in sorts according to their rank. Turns out that this system has a unique set of characteristics: it has similar normalization behaviour as STLC (so complexity of normalization is TOWER-complete) while at the same time allowing non-uniform representation of arbitrary elementary recursive functions; furthermore, it has a nontrivial set-theoretical model in ZF+GCH and allows to carry out nontrivial proof developments. Base normalization proof is surprisingly simple (modulo general metatheoretical properties of pure type systems) and has been fully formalized in vanilla Rocq (formerly Coq) without additional axioms.

0.30

Taming Cyclic Dependencies in Stabilization-Based String Constraint Solving

cs.LO ·2026-10-18 ·Siyi Wu, Bican Xia
Abstract

Word equations with cyclic variable dependencies pose a challenge to SMT string solvers. In stabilization-based procedures, such dependencies can trigger repeated refinement of regular-language assignments without reaching stability. We present Z3-Noodler-C, an extension of Z3-Noodler, combining equality-aware stabilization with dedicated loop reasoning. We introduce equality predicates to retain relevant equations as single stability obligations, and extend stabilization with equality-aware substitution and adapted concatenation-eliminating refinement to expose cyclic dependencies as overlapping variable bindings. Dedicated loop reasoning rules then exploit the periodic structure of these bindings, avoiding certain infinite derivations. We evaluate Z3-Noodler-C against Z3-Noodler, cvc5 and OSTRICH2. On 64 systematically generated, syntactically simple instances involving cyclic dependencies and regular constraints, our solver solves all instances, compared with only 20 solved for the best baseline. On equation-oriented instances from SMT-LIB'25, Z3-Noodler-C achieves the highest coverage, solving 20 more instances than Z3-Noodler while maintaining competitive aggregate CPU time.

0.30

Emulating DGEMM and ZGEMM with INT8 Matrix Engines on Intel Arc GPUs: Ozaki Scheme II on the Xe2 Architecture

cs.MS cs.DCmath.NA ·2026-10-18 ·Rostislavs Rostovskis, Aleksandrs Platonenko
code/data
Abstract · 21 pages, 5 figures, 17 tables. Code: https://github.com/MacraMacra/OS2-Xe

Double-precision matrix multiplication, real (DGEMM) and complex (ZGEMM), dominates the run time of dense linear algebra and electronic-structure codes. On affordable GPUs, however, FP64 throughput is a small and shrinking fraction of the throughput of the low-precision matrix engines. Ozaki scheme II closes this gap by reducing a floating-point matrix product to a handful of exact INT8 products followed by a Chinese-remainder reconstruction; the number of products sets the accuracy. We present what is, to our knowledge, the first implementation of Ozaki scheme II for Intel GPUs. Written in SYCL on top of oneMKL, it runs entirely on the device and exposes the full BLAS dgemm/zgemm interface. Making the scheme fast on the Xe2 (Battlemage) architecture required keeping FP64 arithmetic out of all memory-bound kernels, a branch-free residue computation, an FP32 evaluation of the CRT quotient with a proven safety margin, 16-bit limb arithmetic in the reconstruction, and work-arounds for performance cliffs of the vendor INT8 GEMM. On an Intel Arc Pro B60 (24 GB) the emulated DGEMM reaches 17.3, 12.1 and 9.3 TFLOP/s with $s=8$, 12 and 16 moduli at $N=8192$, and up to 21.1 TFLOP/s overall, i.e. 12-28x the native FP64 DGEMM of the same board. The emulated ZGEMM, which uses the 3M form exactly in integer arithmetic, reaches up to 27.4 TFLOP/s and is 13.7-17.8x faster than native ZGEMM for $N\ge4096$ at $s=16$. With $s\ge16$ the results are correctly rounded for well-scaled inputs and, using an extended-width residue path ($s=20$-24), also for inputs with wide exponent ranges, whereas native DGEMM rounds only 2-13% of the entries correctly. We also report worst-case accuracy, phase breakdowns, energy, an ablation study, a comparison with Ozaki scheme I, and a validated performance model.

0.29

Studying proof data with proof theory: the case of AI-generated mathematics

cs.LO math.LO ·2026-10-18 ·Safal Raman Aryal
Abstract

I present a method for analyzing large collections of formally verified mathematical proofs (e.g. OpenAI's) grounded in some historical and philosophical musings on mathematics in the age of AI. Many of the quantities involved, such as the size of a proof term or the number of redundant steps, are already computed by work on improving automated provers, where they serve as targets. My aim is to turn them into measurements of proofs themselves: compared against a reference library and against each proof's own cut-free form, checked against theorems they must obey, and controlled for how the proofs were compiled. With these instruments, I measure how (much) machine-generated mathematics diverges from human mathematics, and find that AI-written formal proofs realize far less of the logical compression available to them than human mathematicians do (31% against 73% in mathlib, and 55% against 86% on proofs that call no automation). These figures are broadly in line with reports from those who build provers and proof-improvement tools that AI-written proofs are often longer than they need to be.

0.24

Structure-Aware Equivalence for Evaluating Autoformalized Statements

cs.PL ·2026-10-18 ·Yi Yu, Yu-Hsuan Wu, Guang Lin, Yuhui Zhang, Yuxiang Peng
benchmarkhot direction
Abstract

Evaluating autoformalized mathematical statements requires determining whether a candidate preserves the mathematical claim intended by a trusted reference. Existing methods approach this problem from complementary directions: reference-free learned evaluators estimate semantic agreement but are not constrained to a formal equivalence relation, while reference-based proof methods typically employ whole-statement logical equivalence to provide formal certificates, which is too coarse for translation fidelity and may identify mathematically unrelated statements. We introduce Structure-Aware Equivalence (SAE), a role-preserving formal equivalence relation between syntactic and whole-statement logical equivalence that allows formally justified reformulations while preserving translation-relevant mathematical structure. We operationalize SAE in the Autoformalization Correctness Evaluator (ACE), which provides certified equivalence or inequivalence judgments and reports Inconclusive otherwise. On ProofNetVerif, a benchmark introduced with BEq+, ACE retains substantial equivalent-case coverage while remaining highly conservative in certifying disputed or inequivalent pairs. On targeted diagnostics, BEq+ accepts 94-96% of unrelated but whole-statement logically equivalent pairs, while Mathesis LeanScorer accepts 42% of localized meaning-changing perturbations; ACE accepts none of either set as equivalent.

0.23

KernelPack-JAX: a meshless differentiable GPU framework for node generation and forward and inverse PDE problems

cs.MS math.NA ·2026-10-18 ·Alberto Cattaneo, M. Keith Ballard, Robert M. Kirby, Grady B. Wright, Varun Shankar
Abstract

We present KernelPack-JAX, a meshless, differentiable, GPU framework in JAX for solving forward and inverse PDE problems in domains whose boundaries, including internal obstacles, are given only by samples of their smooth pieces. Our node generator is based on a new stacked representation of piecewise-smooth boundaries, in which a parametric partition of unity of each smooth piece induces the supports of a curl-free implicit partition of unity of the whole boundary. This representation keeps the edges and corners of the boundary and classifies every point of space without a global fit. Boundary and interior nodes are then selected by a parallel greedy independent-set algorithm that reproduces sequential Poisson-disk sampling exactly. We discretize elliptic and parabolic problems with radial basis function-generated finite differences (RBF-FD) on stencils selected by a new visibility heuristic to avoid crossing the boundary, with ghost nodes for the boundary conditions. In addition, our framework exposes a GPU sparse direct solver to JAX as a differentiable linear solve, so that Jacobians of time-dependent problems cost a small multiple of a forward solve. We derive error estimates that separate consistency from stability and bound the parameter error by the forward error. We demonstrate high-order convergence for Darcy flow, reaction-diffusion, and advection-diffusion in 2D and 3D domains, including one with an internal obstacle, and show that the recovered parameters largely follow the forward errors. We also present timings for every stage.

0.20

From Software History to Executable Regression Tasks: A System for Evaluating Software Agents

cs.SE ·2026-10-18 ·Ruikai Shi, Zhengran Zeng, Rui Xie, Shikun Zhang, Wei Ye
Abstract · 22 pages

As software agents increasingly generate and modify code, evaluating their ability to review changes and discover defects becomes important. Such evaluation needs executable project environments for investigation and reliable defect references supported by reproducible behavior. Existing code review benchmarks provide limited support for this combination. We present a system that constructs executable regression instances from historical introducing--fixing commit pairs and developer tests. The system extracts candidate tests from fixing-side changes, then uses constrained agents to transplant them and configure version-specific environments while preserving production code and test expectations. Independent replay and evidence-based review verify whether tests that fail on the introducing version and pass on the fixing version support the proposed regression relationship. Accepted instances retain both versions' replay recipes, validated tests, and execution evidence. Repair-supported analysis and human review add root-cause locations, tested repairs, and causal explanations. These instances support bug detection and localization in code review, and enable bug repair and regression-test generation on the same defects. The system constructs 172 validated instances from 522 candidates across 11 Python projects. We evaluate three software agents with ten models and additional reasoning-effort settings on 83 manually selected, enriched instances across the three tasks. Precise root-line localization remains challenging, while baseline combinations achieve repair and regression-test generation success rates of up to 68.7% and 77.1%, respectively. Leading configurations differ by task, and higher task performance does not necessarily yield greater API-cost efficiency. Our approach provides reusable executable evidence for studying agents' defect discovery, repair, and testing capabilities.

0.17

On the Complexity of Level-Synchronized Tree Automata for Quantum Circuit Verification

cs.FL quant-ph ·2026-10-18 ·Michael Blondin, Samy Chady Khoumsi-Kasmi, Guillermo A. Pérez, Marnix Suilen
theory
Abstract

Reliable quantum computation requires methods for verifying quantum circuits. Level-synchronized tree automata (LSTAs) were recently introduced to represent sets of quantum states for this purpose. We study LSTAs along two axes relevant to verification: the complexity of their decision problems and the size of their representations under circuit operations. We prove that acyclic LSTAs have NP-complete nonemptiness and $Π_2^P$-complete language inclusion, and that general language inclusion is EXPSPACE-complete. We further prove inherent quadratic representation growth under individual quantum gates in a fixed qubit order, and give an adjacent-qubit reordering construction that preserves level synchronization. These results clarify the scope of applicability of the automata-theoretic toolbox for quantum circuit verification.

0.13

Noise-Awareness in Quantum Software Testing: An Empirical Landscape Analysis

cs.SE quant-ph ·2026-10-18 ·Mir Mohammad Yousuf, Shabir Ahmad Sofi, Bisma Majid
Abstract · 25 pages, 7 figures, 10 tables, 1 algorithm

Noisy Intermediate-Scale Quantum (NISQ) hardware has transformed quantum computing into a practical experimental platform while introducing fundamental challenges associated with hardware noise, decoherence, and probabilistic execution. Although quantum software testing (QST) has rapidly emerged as an active research area, it remains unclear to what extent existing testing methodologies explicitly engage with noisy and hardware-constrained execution environments. This paper presents a systematic empirical landscape analysis of noise-awareness in QST research through a deterministic, taxonomy-guided analytical framework applied to 61 papers published between 2018 and 2025 across eight testing paradigms, contrasting how contributions are framed with how they are evaluated. Among the analyzed papers, 47 (77.0%) are noise-oblivious and only 14 (23.0%) explicitly engage with noise-related execution concerns, while only 11 (18.0%) report evaluation on physical hardware or a noisy simulator. Test generation, the largest paradigm, exhibits a noise-awareness rate of only 14.3%, and debugging and repair 18.2%; coverage-based, metamorphic, and differential testing contain no noise-aware contributions. Noise-aware methodologies emerged late and remain outnumbered in all analyzed years. Framing and evaluation are strongly associated (odds ratio 40.5, p < 0.001) yet diverge for individual papers: 5 of the 14 noise-aware papers evaluate only on ideal simulators. Inferential statistical tests are reported in only 16.4% of papers. The classification framework was validated against a blinded, human-coded reference sample, achieving almost perfect inter-rater agreement on testing-technique classification and substantial agreement on noise-awareness. The framework provides a reusable empirical basis for longitudinal analysis of hardware-conscious testing practice.

0.09

Theory & Algorithms 24

When Are Monomial Symmetric Polynomials Hard?

cs.CC cs.SCmath.CO ·2026-10-18 ·Markus Bläser, Radu Curticapean, Sagnik Dutta, Gorav Jindal
top venue
Abstract · 20 pages, 3 figures. Accepted at SODA 2027

For a partition $\boldsymbolλ = (λ_1,\ldots,λ_r)$ and $n\ge r$, the monomial symmetric polynomial is defined as $$m_{\boldsymbolλ}(x_1,\ldots,x_n) := \sum x_{i_1}^{λ_1} x_{i_2}^{λ_2} \cdots x_{i_r}^{λ_r},$$ where the sum ranges over all distinct monomials obtained by assigning the parts of $\boldsymbolλ$ to pairwise distinct variables. We establish an almost-complete dichotomy for the complexity of the monomial symmetric polynomials $m_{\boldsymbolλ}$: If a family of partitions contains polynomially many distinct parts, then the corresponding monomial symmetric polynomials (on sufficiently many variables) are VNP-complete. If the number of distinct parts is constant, the family admits polynomial-size circuits. Thus the only remaining gap concerns the intermediate regime where the number of distinct parts grows subpolynomially. As part of our proof, we use combinatorial properties of matching polynomials to establish VNP-hardness of the perfect matching polynomial for blow-up graphs; these are graphs with non-generic edge-weights that are obtained by "blowing up" a smaller graph. We reduce these matching polynomials to monomial-symmetric polynomials by combining classical techniques from algebraic complexity theory with structural properties of Sidon sets.

0.73

Approximating Independent Sets Using Additive Combinatorics

cs.DS cs.CCmath.CO ·2026-10-18 ·Shachar Lovett
published
Abstract

We give a deterministic quasi-polynomial time algorithm for approximating the largest independent set in a graph. We show that if an $N$-vertex graph contains an independent set of size $2^{-t} N$, then we can find an independent set of size $2^t$ in quasi-polynomial time, for all $t$ up to a constant factor of $(\log N)^{1/3}$. This gives an approximation factor of $N / \exp((\log N)^{1/3})$, which has the same shape as known hardness results for quasi-polynomial time algorithms, up to a different exponent of the logarithm. The algorithm builds on a structural principle: a graph containing a large independent set must also contain a large sparse structured set, and such a set can be found efficiently. A greedy algorithm then extracts an independent set from it. Our proof draws on tools in additive combinatorics, in particular spreadness and sifting, to obtain the combinatorial structural results. We then convert these to efficient algorithms using pseudo-marginals obtained by semidefinite programming.

0.60

Rapid Mixing for Sampling Colorings of Random Graphs up to Tree Uniqueness Threshold

cs.DM math.PRcs.DS ·2026-10-18 ·Zhidan Li, Kuan Yang
theory
Abstract · 52 pages

We study the mixing time of the single-site Glauber dynamics for sampling proper $q$-colorings on sparse Erdős-Rényi random graphs $\mathcal{G}(n, d/n)$. We prove that, whenever $q \ge (1 + o_d(1))d$, Glauber dynamics mixes in polynomial time with high probability over $\mathcal{G}(n, d/n)$ for all sufficiently large integers $n$. This improves the previous bound $q = (α^* + o_d(1))d$ where $α^* = 1.7632...$ is the unique solution of $x = \exp(1/x)$ on $[1, \infty)$ by Efthymiou, Hayes, Štefankovič, and Vigoda (SODA, 2018) and asymptotically matches the uniqueness threshold on infinite regular trees. Our proof has three main ingredients. First we establish logarithmic bounds on the spectral and coupling independence constants of the Gibbs distribution by extending the tree analysis of Chen, Liu, Mani, and Moitra (J. ACM, 2026). Second, we extend the entropy-factorization framework for two-spin systems on random graphs, developed by Bezáková, Galanis, Goldberg, and Štefankovič (TALG, 2024), to proper colorings by introducing a buffer around high-degree vertices. Finally, we compare the resulting heat-bath block dynamics with the single-site Glauber dynamics.

0.49

Range Queries for Permutation Pattern Matching: Separations, Equivalences, and Applications

cs.DS ·2026-10-18 ·Omri Ben-Eliezer, Tomer Even
top venue
Abstract · 48 pages. To appear in SODA 2027. Abstract shortened to meet arXiv requirements

The problem of Permutation Pattern Matching (PPM) asks, given a permutation $π:[n]\to[n]$, viewed as the point set $P=\{(i,π(i)):i\in[n]\}$, and a pattern $σ$ of small length $k$, whether $P$ contains an order-isomorphic copy of $σ$. The PPM problem and its (exact and approximate) counting variants have been studied extensively in the literature. We consider a preprocessing-query version: after preprocessing $P$, one must determine whether a query $P'\subseteq P$ contains $σ$, where $P'$ is obtained by intersecting $P$ with a geometric shape (e.g., an axis-aligned rectangle or $k$ such rectangles). This naturally generalizes the classical range query problem in computational geometry, which corresponds to the case $|σ|=1$. We extend the study of this problem from patterns of length $2$ to patterns of length $3$ (and in some cases beyond), and prove new fine-grained equivalences, separations, and applications: (1) Single rectangle: For all patterns of length $3$, detection can be done with $\tilde{O}(n)$ preprocessing and $\tilde{O}(1)$ query time. The proof goes through a new tool of independent interest: reductions from rectangle queries to a simpler family of slab queries. (2) Multiple rectangles: For $σ=132$, detection is equivalent to the online set disjointness problem; we also show connections to online $k$-set disjointness. To our knowledge, this is the first equivalence between a PPM-type problem and online set disjointness, a problem playing a fundamental role in fine-grained complexity. (3) Counting: As applications, we obtain an $O(n^{1.725})$ time algorithm for exact counting of $5$-patterns based on the rectangle-to-slab reduction, improving upon [Beniamini and Lavee, ICALP 2025], and an $\tilde{O}(n^{7/4})$ approximate counting algorithm for $6$-patterns, the first to break the $O(n^2)$ barrier for $6$-patterns.

0.46

A Self-Coupling Gaussian Extremal Inequality for Information-Constrained Optimal Transport with Applications to Rate-Distortion-Perception

cs.IT ·2026-10-18 ·Jiayang Zou, Jun Chen, Ayfer Özgür, Jia Wang
theory
Abstract · 31 pages. Submitted to IEEE Transactions on Information Theory

How much mutual information is needed to achieve a target correlation between two copies of a prescribed distribution? We prove a sharp Gaussian extremal inequality for this self-coupling problem: among distributions with a fixed mean and covariance, the Gaussian has the greatest minimum information cost for meeting a positive semidefinite cross-covariance target. The inequality is strict for non-Gaussian distributions when the target is positive definite. The proof uses a midpoint gluing construction. Two Ornstein--Uhlenbeck half-bridges, each started from the prescribed endpoint law, meet at a common intermediate distribution and are joined by conditional gluing. Changing the Gaussian reference then turns the resulting relative-entropy bound into the desired extremal inequality. As a first application, scaling one endpoint of the self-coupling establishes Gaussian extremality for rate-distortion-perception with unlimited common randomness, under quadratic distortion and squared Wasserstein-2 perception measures. We also establish an analogous result for finite common randomness, where the problem involves two mutual information constraints associated with an auxiliary random variable. In that setting, we retain an intermediate observation on one half-bridge and balance the two information budgets simultaneously using the Poincaré-Miranda theorem.

0.36

Pre-Query Retention with Bounded Bit and Word Retrieval

cs.IT cs.AI ·2026-10-18 ·Pranav Singh, Prashant Singh
Abstract · 24 pages, 4 figures. Complete proofs and experimental details included. Both authors contributed equally

An agent can retain information before receiving a future question and later retrieve evidence from an archive. Evaluating this arrangement requires accounting for both retained information and the width of each retrieval response. Decision-tree capacity bounds describe individual sets of distinguishable records, but a memory writer must encode every possible record before the query is known. We study fixed-length retention beside an immutable binary archive, with a worst-case limit on raw bit or word reads. We characterize feasible one-bit-read memory cells and construct encoders that partition the complete record space. Supporting pairwise exclusive disjunction and conjunction with one bit read is equivalent to supporting every Boolean query. For multiple reads, we combine explicit encoders with decision-tree counting and an entropy converse. The word-read bounds differ by at most two retained bits when the read budget is at most half the archive length. Under an explicit condition on word width and archive length, raw-word caching attains the exact optimum among all writers. A related condition preserves this optimum under positive average error. Exhaustive checks validate the implementations. In a matched dialogue-agent study, a categorical sufficient statistic preserves source decisions with less storage and lower decoding latency than the universal encoder. A catalogue-specific index reduces private state further and requires additional shared storage. The results characterize how response width and future query structure determine retention requirements.

0.33

Searching subgraphs

cs.DS ·2026-10-18 ·Dariusz Dereniowski, Boting Yang
hot direction
Abstract

We consider graph search models, in which an invisible, omniscient and fast intruder can hide anywhere in the graph, and the goal is to design a \emph{search strategy} composed of a sequence of movements of \emph{searchers} that guarantees the capture of the intruder regardless of its actions. Our point of departure is the connected search model in which the cleared subgraph is always connected during the search process. In this paper, we initiate the study of the subgraph searching problems which arise from the following question: what happens if the number of searchers available is not sufficient to clear a given graph $G$? We propose an approach in which one wants to clear in a connected way as large subgraph as possible with the given $k$ searchers. In this context we study two problems: one problem, called the \emph{subgraph search}, is to find the maximum-sized subgraph whose search number is at most $k$, and the other one, called the \emph{protected subgraph search}, is to perform a partial search on $G$ leading to as many cleared edges as possible. Informally speaking, the difference lies in the fact that in the latter a recontamination may come from an edge never to be cleared and hence the partial search strategy needs to avoid such recontaminations. Having proved that both problems are computationally hard in general, we give efficient algorithms for $n$-vertex trees: an $\bigo(\min\{kn, n \log n\})$ time algorithm for the subgraph search and an $\bigo(\min\{2^{6k}kn, n^{8.571} \log n\})$ time algorithm for the protected subgraph search.

0.32

MMS Allocation for Chores with Item Disposals

cs.GT ·2026-10-18 ·Shinuo Ma, Xiaowei Wu
benchmarktheory
Abstract

Web platforms organize collective work, including content review and community maintenance, that imposes heterogeneous effort costs on participants. Fairly distributing these burdens while leaving as few tasks unassigned as possible is a natural objective for the design of such platforms. We study this through the maximin share (MMS) allocation problem with chore disposals under nonnegative additive costs. Our goal is to minimize the number of unassigned chores while guaranteeing every agent her MMS benchmark from the original instance. We prove that, for every instance with $n \ge 4$ agents, at most $n-3$ chore disposals suffice. We improve the bound of $2m/11+n$ disposals established by Barman et al. (2025), where $m$ is the number of chores, and eliminate the dependence on $m$. For four agents, we obtain a tight bound: one disposal always suffices, whereas known instances admit no complete MMS allocation (Ezra and Garbuz, 2026). We also prove the same result for goods: $n-3$ duplications suffice to ensure MMS.

0.31

Dimension-free comparison inequalities between $f$-divergences for Gaussian location mixtures

cs.IT math.STmath.PRstat.ML ·2026-10-18 ·Jeonghwan Lee, Cong Ma
Abstract · 21 pages; Comments would be appreciated!

We establish dimension-free comparison inequalities between the Kullback-Leibler (KL) and $χ^2$-divergences and the squared Hellinger distance for Gaussian location mixtures with compactly supported mixing distributions. We show that both divergences are bounded by the squared Hellinger distance up to constants depending only on the radius of the support, independently of the ambient dimension and the number of mixture components. These results resolve an open question raised by \cite{jia2023entropic} and strengthen existing comparison inequalities. Our proof relies on a Gaussian interpolation argument that relates the derivatives of these divergences to differences between posterior means.

0.29

Formally Verified Autoresearch for Analyzing the Expressivity of Multi-Head Attention

cs.CC cs.AIcs.LG ·2026-10-18 ·Karthik Viswanathan, Plamen Dimitrov, Stefano Rocca, Sarayu Sirikonda
publishedtheory
Abstract · 25 pages, 1 figure. Oral presentation at the AutoMLR Workshop at NeurIPS 2026

Multi-head attention is a central component of the transformer architecture, yet how its expressivity depends on the number of heads remains poorly understood. We study a one-layer, attention-only transformer with additive positional embeddings and define the head complexity $H^\ast(f)$ of a Boolean function $f$ as the minimum number of attention heads required to represent it. We derive an exact linear-fractional normal form for head complexity, prove a general threshold-degree lower bound, and give an explicit ten-bit function for which this bound is strict. For symmetric Boolean functions, we independently show that head complexity equals the number of times the function changes value between consecutive Hamming-weight levels, even when positional embeddings are allowed. This agrees with the concurrent characterization of Hsu and Xu (arXiv:2608.04243) in a model without positional embeddings. The symmetric lower bound combines our general threshold-degree bound with the classical result of Aspnes et al. (1994). These results were discovered and formalized by a frontier model through conjecture generation, theorem proving, and machine verification, while human researchers asked the high-level questions and reviewed the generated theorems. We make the Lean formalization and research traces publicly available to support agent-assisted theorem proving and formal verification in mechanistic interpretability. This work provides a proof of concept for using the capabilities of frontier models, grounded by formal verification, to develop an understanding of transformer architectures.

0.29

Applying Changes to LDPC Codes by Using 3-Induced Matchings and Eigenvalues

cs.IT math.CO ·2026-10-18 ·Alireza Eshraghi Dehaghani, Sara Saeedi Madani, Dariush Kiani
Abstract

The 2-switch operation for reducing the number of short cycles in Tanner graphs of LDPC codes has been utilized in some works. If this operation is appropriately applied to a 2-induced matching in a bipartite graph, then it surely preserves the bipartiteness of the graph. With a 2-induced matching, we have only one choice for applying this type of change to the graph. Using 3-induced matchings creates more potential and flexibility to apply certain changes to the Tanner graph. We present a necessary and sufficient condition for the existence of a 3-induced matching in bipartite graphs with girth at least 6. A 3-induced matching helps us to have more choices for applying a change to a bipartite graph while surely preserving the bipartiteness of the graph and keeping the degree sequences unchanged. Among all changes that we consider, we select the best change for reducing the number of cycles of length 6. The eigenvalues of subgraphs are used to reach a 3-induced matching, and a condition in some steps can be checked to ensure that the subgraph has at least one 3-induced matching. We present some numerical results regarding the steps at which the condition that guarantees the existence of a 3-induced matching is checked.

0.27

Degree Stability of Parallel Greedy Matching in Random Graphs

cs.DM cs.DSmath.PR ·2026-10-18 ·Vishvajeet N, Shrinidhi Teganahally Sridhara
hot direction
Abstract · 49 pages

We study a natural synchronous greedy matching process in Erdős-Rényi random graphs $\mathbb{G}(n,p)$ and give a rigorous analysis of this parallel stochastic dynamics. The process evolves as follows: in each round, every unmatched vertex proposes to a uniformly random neighbour, and mutually proposing pairs are matched and removed from the graph, after which the process repeats on the residual graph. We consider the density regime $p=Θ(\log^{1+ε} n/n)$ for any constant $ε>0$, which lies just above the threshold for the existence of a perfect matching. We show that, with high probability, for $Θ(\log^{1+ε} n)$ rounds, every unmatched vertex keeps a constant fraction of its initial degree in the residual graph, simultaneously for all vertices and all rounds. By then, the process has matched a constant fraction of the vertices. Our analysis tracks fine-grained invariants that control the evolution of individual vertex degrees, rather than directly tracking the global degree distribution. At the local level, we analyze the per-round degree loss of a vertex via a neighbourhood-exposure martingale. We show that structural sparsity of $\mathbb{G}(n,p)$ yields $O(1)$ variance and thus the per-round degree loss is strongly concentrated around its conditional expectation. At the global level, these local concentration bounds control the increments of a martingale tracking cumulative degree loss over many rounds, allowing us to show that vertex degrees remain stable for $Ω(\log^{1+ε} n)$ rounds.

0.27

No Escape from the Quadratic Barrier: Adaptive Attacks on Linear $\ell_2$ Sketches

cs.DS cs.LG ·2026-10-18 ·Edith Cohen, Uri Stemmer
Abstract

Linear sketches estimate properties of high-dimensional inputs from a fixed low-dimensional projection. At constant accuracy, randomized Euclidean-norm sketches, notably the Johnson-Lindenstrauss transform, support exponentially many nonadaptive queries in sketch dimension $k$. For adaptive black-box queries, known privacy-based defenses guarantee $\widetildeΩ(k^2)$ and it remained open whether stronger guarantees were possible. Prior attacks used more queries, restricted the responder, or provided weaker failure guarantees. We show that the quadratic guarantee is optimal up to logarithmic factors, even with a computationally unbounded responder that knows the matrix, attack algorithm, history, and query distributions and may change its randomized rule after every query. For every fixed hidden $A\in\mathbb{R}^{k\times n}$ with $n>40k$ and target accuracy $α\in (0,1]$, our black-box attack uses $O(k+α^2k^2\log(2+α^2k))$ queries in $\widetilde{O}(1+α^2k)$ adaptive rounds to force, with constant probability, a round in which a constant fraction of answers have $Ω(α)$ relative error. This fraction can be taken arbitrarily close to one half. Our results clarify the limits of defending a fixed compressive linear representation by adapting the prediction rule.

0.25

A Faster Strongly Polynomial Time Algorithm for Submodular Function Minimization

cs.DS ·2026-10-18 ·Alina Ene, Huy L. Nguyen
Abstract

We give a strongly polynomial time algorithm for submodular function minimization with overall running time $\widetilde{O}(n^{2}\mathsf{EO}+n^{3})$ time, where $\mathsf{EO}$ is the running time of a single evaluation of the function. Our algorithm improves upon the previously fastest strongly polynomial algorithms of Lee et al. (FOCS 2015), Dadush et al. (SODA 2018), and Jiang et al. (SODA 2024) that run in time $\widetilde{O}(n^{3}\mathsf{EO}+n^{4})$. Similarly to prior works, our work combines a cutting plane algorithm with the combinatorial arc-finding framework of Fleischer et al. (JACM 2001). We achieve our faster running time by developing an algorithm that runs the cutting plane method end to end, without restarting after discovering arcs and contracting strongly connected components.

0.24

Convergence of Denoising Diffusion Probabilistic Models: A Rényi Divergence Perspective

cs.IT ·2026-10-18 ·Wenliang Liu, Santiago Mazuelas, Guanding Yu, Renjie Liao, Lele Wang +1 more
Abstract

Denoising diffusion probabilistic models (DDPMs) have emerged as a powerful framework for generative modeling by transforming standard Gaussian inputs into outputs that approximate a target distribution. While DDPMs have been proved to converge in total variation, KL divergence, and Wasserstein distance, these discrepancy measures share a fundamental limitation of being insensitive to distributional mismatch in low-probability regions. To address this, we adopt the stronger Rényi divergence as the discrepancy measure. Leveraging a Girsanov-type change-of-measure argument, we establish the first Rényi characterization of DDPMs to the best of our knowledge, showing that the total error decomposes into a mixing error, a discretization error, and a score estimation error. We assume the target distribution is sub-Gaussian and characterize the score error via the cumulant generating function, relaxing both the log-Sobolev-inequality and $L^\infty$-score-error conditions used in prior Rényi divergence analyses of inexact Langevin Monte Carlo, a score-based sampling algorithm closely related to DDPM. Our results also establish an iteration complexity for achieving $ε$-accuracy in $α$-Rényi divergence that is asymptotically proportional to $α^3/ε$. Finally, by extending our analysis, we establish the first Rényi divergence convergence guarantee for critically damped DDPMs, a kinetic variant governed by critically damped Langevin dynamics. Similar to existing characterizations of critically damped DDPMs, this result shares the structural scaling of standard DDPMs, differing by at most a logarithmic factor in the mixing error exponent, and supports the conjecture that critically damped DDPMs do not yield a theoretical improvement.

0.21

Strategy recovery for Mean-Payoff Games is as hard as solving them from scratch

cs.GT cs.DS ·2026-10-18 ·Dani Dorfman, Uri Zwick
published
Abstract · To appear in SOSA 2027

Mean-Payoff Games (MPGs) form an interesting family of deterministic, infinite-duration, zero-sum, turn-based games played by two players on a finite weighted directed graph. The goal of the players is to minimize, respectively maximize, the $\textit{limiting average}$ of the costs of the edges on an infinite path constructed by the two players. The complexity of solving MPGs is a long-standing open question. The decision problem associated with these games is known to be in $\text{NP}\cap \text{co-NP}$, and even $\text{UP}\cap \text{co-UP}$, yet the problem is not known to be in $\text{P}$. This is one of the simplest combinatorial problems with such a status. A full solution of an MPG gives the $\textit{value}$ of each vertex of the game, along with a pair of $\textit{optimal strategies}$ for the two players. (Each player has a succinct $\textit{positional}$ optimal strategy.) The problem of finding optimal strategies for the players given the values of all vertices is known as the $\textit{strategy recovery}$ problem. We describe a simple reduction from the problem of solving MPGs to the strategy recovery problem for MPGs. This shows that the strategy recovery problem for MPGs is essentially as hard as solving MPGs from scratch, i.e., without knowing the values. The reduction uses simple $\textit{game transformations}$ and simple $\textit{self-reducibility}$ and $\textit{peeling}$ arguments. This answers a question of Andersson and Miltersen. Mamino obtained a similar result for $\textit{stochastic}$ games with a limiting average objective, but his technique does not extend to deterministic games such as MPGs. We go one step further and show that finding an optimal strategy for one of the players of an MPG remains hard even if an optimal strategy of the second player, and hence all the values, are known.

0.19

Breaking the 1/3 Barrier for Approximate Nash Equilibria in Symmetric Bimatrix Games

cs.GT cs.LG ·2026-10-18 ·Ashkan Soleymani, Patrick Jaillet, Gabriele Farina
published
Abstract · ACM-SIAM Symposium on Discrete Algorithms (SODA), 2027

We give the first unconditional family of polynomial-time algorithms that breaks the $1/3$ approximation barrier for symmetric bimatrix games. For every fixed $δ>0$, an algorithm in this family computes a symmetric $(1-1/\sqrt{2}+δ)$-approximate Nash equilibrium, improving on the previous best guarantee of $(1/3+δ)$ due to Kontogiannis and Spirakis [2011].

0.14

Can Quantum Entanglement Increase the Capacity of a Classical Relay Channel?

cs.IT quant-ph ·2026-10-18 ·Elsa Mizrahi, Uzi Pereg
theory
Abstract

We show positive and negative results on classical relay channels with entanglement assistance, revealing a sharp contrast between strictly causal and causal relay encoding. First, we consider van der Meulen's setting of a unit-delay relay, with strictly causal encoding. We show that the standard cutset upper bound continues to hold with entanglement assistance, and deduce that entanglement does not increase the capacity for several classes of relay channels. Second, we consider a zero-delay relay, where the encoding is causal. We present an example in which entanglement enables rates that exceed the cutset upper bound by at least 12.5%. Namely, entanglement shared between the source and relay significantly increases the capacity of a classical zero-delay relay channel.

0.14

From Formulation to Articulation: Considerations for AI-Assisted Research in Information Theory

cs.IT ·2026-10-18 ·Chao Tian, Krishna Narayanan
Abstract

Recent advances in artificial intelligence have contributed to several information-theoretic results through genuinely different modes: a language model may suggest a proof, participate in an extended mathematical dialogue, coordinate computation, or operate as one component of a search harness. Treating all of these as "AI-discovered mathematics" obscures distinctions that matter for scientific reliability, attribution, and reproducibility. This article examines a set of recent cases through a four-stage pipeline: how the problem was formulated; how AI participated in finding a solution; how the resulting analytical and computational claims were verified; and how machine-generated outputs were translated into mathematical exposition for human understanding. Based on these cases, we recommend that research publications (1) disclose the nature and extent of AI assistance at each stage, (2) report verification of analytical reasoning separately from certification of numerical computations, and (3) place more emphasis on articulating the underlying ideas, techniques, and interpretations.

0.13

Non-Monetary Mechanisms for Stochastic Scheduling via Sampled Pairwise Externality Compensation

cs.GT ·2026-10-18 ·David X. Lin, Chido Onyeze, Giannis Fikioris, Siddhartha Banerjee, Éva Tardos
benchmark
Abstract

How can a shared resource be allocated without money when agents have private, stochastic demands that can be queued for a while, and may behave strategically, or even adversarially? We introduce a paradigm for non-monetary online resource allocation called Sampled Pairwise Externality Compensator (SPEC). Our objective is to guarantee each agent a large fraction of its ideal utility: an individual benchmark corresponding to serving the highest-valued fraction of its demand permitted by an equal share of capacity. We seek guarantees that hold regardless of other agents' behavior, including potentially adversarial and collusive behavior. Our approach builds on the idea of pairwise externality compensation, making it computationally efficient by replacing exact calculations of future externalities with sampled estimates. The central insight is that local mixing, exploited through coupled simulations, allows these estimates to be computed efficiently while preserving robust utility guarantees. In our main technical contribution, we instantiate SPEC in a stochastic scheduling model in which agents receive jobs with private values, unserved jobs persist across periods subject to random expiration, and a single server processes at most one job per period. Our mechanism gives each agent a simple strategy that achieves the optimal worst-case fraction of ideal utility available to a fully informed planner, up to a sublinear additive loss in total utility. We also show that this fraction approaches $1$ in the long-horizon limit as the expiration probability vanishes: sufficiently persistent jobs allow each agent to secure nearly its full ideal utility. The mechanism runs in $O(n^2)$ expected amortized time per period, where $n$ is the number of agents; significantly, the expected amortized computational overhead per period is independent of the time horizon $T$.

0.13

Subexponential Algorithms for Total Search Problems in the Dense Pigeonhole Regime

cs.DS cs.CC ·2026-10-18 ·Jadwiga Czyżewska, Karolina Drabik, Marek Sokołowski, Karol Węgrzycki
theory
Abstract

The class $\mathsf{PPP}$ captures fundamental total search problems, for which a solution is guaranteed to exist by the pigeonhole principle and the main challenge is to find it. We ask whether problems in this class admit significantly faster algorithms in the \emph{dense regime}. More concretly, we study \emph{binary collision Short Integer Solution} (SIS): given a matrix $A\in\mathbb{Z}_q^{n\times m}$ with $m>n\log q$, the goal is to find distinct $\underline{\mathbf{x}},\underline{\mathbf{y}}\in\{0,1\}^m$ with $A\underline{\mathbf{x}}=A\underline{\mathbf{y}}\pmod q$. Brute-force enumeration can be used to find a collision in $q^{\mathcal{O}(n)} \operatorname{poly}(|I|)$ time, and no significantly faster algorithm for that problem is known. We present an algorithm running in time $q^{o(n)}\cdot\operatorname{poly}(|I|)$ whenever the density parameter $δ:= m/(n\log q)$ is $ω(1)$.\footnote{All logarithms are in base $2$, and $|I|$ is the binary input length.} Building on reductions of Sotiraki et al.~(FOCS~2018), we obtain corresponding algorithms for dense variants of weak constrained SIS, Equal Vector Sums, Minkowski, and Blichfeldt. Our algorithm adapts techniques of Bhati et al.~(FOCS~2026), who gave a subexponential algorithm for Equal Sums in the dense regime and proved a tight conditional lower bound. Their conditional lower bound easily transfers to SIS and shows that the running time of our algorithms is essentially optimal.

0.11

LiftDiff: Guided Discrete Diffusion for Quantum Code Construction

cs.IT quant-phcs.LG ·2026-10-18 ·Yotam Peled, Eliya Nachmani
Abstract

The realization of fault-tolerant quantum computation relies on Quantum Error-Correcting Codes (QECC). Of particular interest are codes that combine a high encoding rate with sufficient distance. Existing algebraic constructions, such as lifted product (LP) codes, build a large quantum low-density parity-check (qLDPC) code from a classical base matrix. In this work we introduce LiftDiff, a generative approach to the discovery of QECC. A discrete denoising diffusion model is trained to generate base matrices over the group algebra $\mathbb{F}_2[\mathbb{Z}_\ell]$, which are then expanded into LP codes. The reverse diffusion process is guided by an energy function that evaluates quantum code properties, including girth, degree, sparsity and decoding performance, steering the sampler toward matrices that yield LP codes with favorable parameters. Our method discovers codes with parameters such as $[[169,13,11]]$, $[[143,15,9]]$, and $[[260,24,10]]$, including a $[[247,19,14]]$ code that achieves the highest ${kd^2}/{n}$ among the baselines. Moreover, several discovered codes outperform the Gross code across a broad range of the tested physical error rates under both code-capacity and circuit-level noise.

0.10

When Weak Submodularity Meets Matroids: How Does Standard Greedy Behave?

cs.DM math.OCcs.LG ·2026-10-18 ·Qixin Zhang, Shi Fu, Shengminjie Chen, Chao Qian, Dacheng Tao
theory
Abstract · 33 pages

Weakly submodular maximization finds numerous applications in machine learning, data mining and statistics. In particular, under a cardinality constraint, the standard greedy algorithm is known to attain the tight $(1-e^{-γ})$-approximation guarantee, where $γ$ denotes the submodularity ratio. However, when considering general matroid constraints, \emph{how well standard greedy can perform} has remained a longstanding open question. In this paper, we resolve this question by studying the worst-case approximation guarantee $R_{\mathrm{gr}}(γ,r)$ of standard forward greedy, where $r$ is the rank of matroid. Specifically, our first contribution is the following refined lower bound \begin{equation*} R_{gr}(γ,r)\ge \max\left\{ \fracγ{r}, \frac{γ^2} {B_γ+\sqrt{γ(r-1)}}, \frac{γ^{r+1}} {2-γ+γ^{r+1}}, \frac{γ^2} {14\left(B_γ+\sqrt{γ(1-γ)(r-1)}\right)} \right\}, \end{equation*} where $B_γ\triangleq2-γ+γ^2$. This lower bound not only strictly improves upon the previous guarantee of $0.4γ^2/(1+\sqrt{γr})$, but also can recover the classical $1/2$-approximation guarantee at $γ=1$. Subsequently, a corresponding upper bound is established for $R_{\mathrm{gr}}(γ,r)$, namely, \begin{equation}\label{eq_upper_abstract} R_{gr}(γ,r) \le\min\left\{\frac{1}{2},20\max\left\{\fracγ{r}, \frac{γ^2}{B_γ+ \sqrt{γ(1-γ)(r-1)}}\right\}\right\}. \end{equation} For every fixed $γ\in(0,1)$, this equation effectively sharpens the recent $O_γ(1/\log(n-1))$ upper bound to $O_γ(r^{-1/2})$, where $n$ is the size of ground set. Putting all these results together, we eventually can obtain \[R_{gr}(γ,r) = Θ\left( \max\left\{\fracγ{r}, \frac{γ^2} {1+\sqrt{γ(1-γ)r}}\right\} \right).\]

0.09

A Bulow-Klemperer Theorem for Bundling and Improved Approximation Bounds for Simple Auctions

cs.GT econ.TH ·2026-10-18 ·Yichen Huang
theory
Abstract

We study simple auctions for one additive bidder whose values for $m$ items are independent. Write $\mathrm{REV}$ for the optimal revenue, $\mathrm{SREV}$ and $\mathrm{BREV}$ for the optimal revenues from separate selling and grand bundling, and $\mathrm{BSPA}_2$ for the grand-bundle second-price revenue with two independent bidders. For independent regular items, we prove $\mathrm{REV} \le \mathrm{BSPA}_2 \le 2\,\mathrm{BREV}$. Thus, one additional bidder lets a reserve-free second-price auction for the grand bundle match the optimal revenue from one bidder, regardless of the number of items. This gives a Bulow-Klemperer theorem for bundling and resolves an open question of Beyhaghi, Cai, Feng, Li, and Weinberg (EC 2025) on whether constant competition suffices for bundling. The factor-$2$ bundling approximation is also tight. For general independent items, we prove $\mathrm{REV} \le \mathrm{SREV} + \frac{π^2}{6}\mathrm{BREV} \le 2.645 \max\{\mathrm{SREV}, \mathrm{BREV}\}$. This improves the previous approximation factors of $6$ from Babaioff, Immorlica, Lucier, and Weinberg (FOCS 2014, JACM 2020) and $5.2$ from Ma and Simchi-Levi (AISTATS 2021). For the lower bound, we construct an instance in which pricing three disjoint bundles earns more than $2.11\max\{\mathrm{SREV},\mathrm{BREV}\}$, improving the previous lower bound of $2$ from Rubinstein (ITCS 2016). Together, these results narrow the tight approximation ratio to the range $[2.11, 2.645]$.

0.07

Robotics, Control & Signals 38

A Novel OFDM-FMCW Waveform for Low-Complexity Joint Sensing and Communication

eess.SP ·2026-10-18 ·Amir Bouziane, Salah Eddine Zegrar, Huseyin Arslan
publishedhighly cited
Abstract

With the advent of 6th Generation (6G) wireless networks, integrating Orthogonal Frequency Division Multiplexing (OFDM) and Frequency Modulated Continuous Wave (FMCW) waveforms presents a promising solution for simultaneous communication and sensing. While OFDM offers high communication performance, it requires complex transceivers for sensing applications. Conversely, FMCW is easily implemented for sensing but provides low data rates for communication. This letter derives the mathematical relationship between OFDM and FMCW, showing how FMCW can be generated from the diagonal elements of the OFDM modulator (\ie, the Discrete Fourier Transform (DFT) matrix). We propose a novel approach that leverages the DFT matrix to seamlessly integrate these two waveforms, minimizing interference and enabling high data rate communication alongside low-complexity sensing receiver design. Extensive simulations demonstrate the effectiveness of our technique in achieving simultaneous sensing and

0.62

BiDex-100: A Comprehensive Benchmark for Bimanual Dexterous Manipulation Across 100+ Simulation Tasks

cs.RO cs.CVcs.LG ·2026-10-18 ·Yong Wang, Kecheng Zheng, Xing Zhu, Chao Xu, Yujun Shen +1 more
benchmark
Abstract

Bimanual manipulation with dexterous hands has attracted growing attention in research on robotic action models. However, few studies have developed benchmarks specifically tailored for dexterous manipulation, leaving the dexterous manipulation capabilities of existing models insufficiently characterized. In this paper, we present BiDex-100, a simulation benchmark for bimanual dexterous manipulation that combines broad task coverage, multi-dimensional evaluation, and high-DoF hand embodiments. Concretely, the benchmark comprises 106 tasks organized along five capability axes: generalization, instruction following, limited demonstrations, long horizon, and memory. Each task is paired with an out-of-domain variant featuring controlled visual and spatial changes, allowing systematic assessment of robustness. In addition, the benchmark supports two embodiments and provides 8,080 success-validated teleoperated demonstrations, totaling 6.25 million frames. We evaluate seven recent Vision-Language-Action models under a unified protocol, establishing baselines for capability-specific performance, robustness, and data efficiency. The website is available at https://robbyant-research.github.io/BiDex-100/.

0.52

RAIL-5G: A Challenging Real-World Dataset and Benchmark for AI-Based 5G Indoor Localization

eess.SP cs.LG ·2026-10-18 ·Jonas Pirkl, Andreas Eidloth, Maximilian Kasparek, Tobias Feigl, Christopher Mutschler
publishedbenchmarkcode/data
Abstract · To be published in the proceedings of the 16th International Conference on Indoor Positioning and Indoor Navigation (IPIN)

Reliable radio-based indoor localization is essential for industrial automation, yet existing methods are often evaluated in environments that do not fully reflect the severe multipath and non-line-of-sight (NLOS) conditions encountered in real deployments. This limits our understanding of how AI, i.e., modern geometric, unsupervised, and semi-supervised localization methods generalize to industrial 5G settings. We present RAIL-5G, a public real-world 5G NR channel impulse response (CIR) benchmark recorded in an operational railway maintenance facility. RAIL-5G covers an 811m$^2$ indoor area with eight distributed radio units, robot-mounted user equipment (UE) trajectories, high-accuracy iGPS reference positions, and severe metal-induced multipath caused by train compartments and maintenance infrastructure. We benchmark geometric time-difference of arrival (TDoA), supervised fingerprinting, unsupervised channel charting, and self-supervised wireless pretraining methods. Our results highlight the difficulty of RAIL-5G: on the full dataset, TDoA reaches 7.02m mean average error (MAE), the best unsupervised method reaches 2.46m, and supervised fingerprinting achieves 0.52m. Self-supervised pretraining reduces the labeling effort by up to one order of magnitude, but the remaining gap to supervised learning shows that robust AI-based localization under severe multipath remains an open challenge. The dataset and code are available at https://iis.fraunhofer.de/rail-5g.

0.39

Vision as Proprioception: Estimating Humanoid Robot Pose from Video

cs.RO cs.AIcs.CV ·2026-10-18 ·Jiaxi Jiang, Matthias Heyrman, René Zurbrügg, Marco Hutter, Christian Holz
code/data
Abstract · Code: https://github.com/eth-siplab/HumanoidPose

Estimating a humanoid robot's internal state from video has a wide range of real-world applications but remains unexplored. During interactions with a robot, its onboard state information may be unavailable due to privacy, ownership, or communication constraints. Meanwhile, online and AI-generated robot videos provide rich sources of data but typically lack corresponding joint configurations, limiting their direct use in robot learning. To address these issues, we present HumanoidPose, a new framework that treats vision as proprioception to recover full-body joint configurations in world space from video alone. Specifically, we combine pretrained human-pose features with robot kinematics in a unified model that predicts each embodiment's joint configurations. We retarget human motions to sixteen different humanoid embodiments, enabling scalable learning entirely from synthetic data. To improve real-world generalization, we further leverage unlabeled online videos through self-training to develop a 2D robot-pose detector. In our evaluation, HumanoidPose improves joint-angle and joint-position accuracy over the evaluated baselines and reduces joint-angle error by 18-32% relative to the best-performing baseline on each of the three real humanoid platforms. We further show improved global localization and generalization to unseen embodiments. HumanoidPose enables visual understanding of other robots during interaction and extraction of robot motion from online and generated videos.

0.38

Compile Once, Run Cheap: Frozen, Portable Skills for Embodied Agents, and What Recursive Self-Improvement Adds

cs.RO cs.AIcs.LG ·2026-10-18 ·Jingwen Bai, Guochao Bai, Canbo Zhang, Shuangqi Li, Minchao Hu +2 more
code/data
Abstract · 28 pages, 4 figures, 13 tables. Code and data: https://github.com/Orillusion/embodied-skill-rsi

Embodied AI lacks the data-rich pretraining that gave language models their agents and skills, so new manipulation capability costs demonstrations per task or a frontier model in every episode. We study a third route: a strong general model acts only as an offline author that writes, runs, diagnoses and repairs a manipulation skill in simulation until it passes an exam on layouts drawn after freezing; the frozen skill then runs as code alone or under a cheap caller that reads only its description. We call this skill-level recursive self-improvement Skill-RSI and specify a compile protocol (isolated, audited author; hidden validation layouts; no-feedback and inherit-only controls; tamper-evident ledger) and a reuse protocol that runs the identical files with executors from another model family. On 7 RoboTwin 2.0 tasks with cluttered tables, Claude Code compiled 4 tasks into 7 frozen skills, 5 from a first draft needing only qualification, at USD 4.29-8.22 of author cost each (USD 55 per skill counting the 13 failed campaigns). The other three tasks did not compile in 18 campaigns, including five with GPT-6 and a pre-registered 150-layout run: on the hardest task every campaign scored 78-95 of 100 but had 3-16 safety aborts, failing the gate of 95% success with zero aborts; a 100-episode exam bounds an abort rate only to 3.6%. On hidden layouts, repair that inherits the previous version beat from-scratch rewriting in 8 of 8 controlled campaigns (sign test p = 0.004); an inherit-only lineage attributes most of the gap to inheritance, with development feedback adding a little (3 of 3, p = 0.125). Write-once campaigns found no detectable first-draft effect of practice (minimum detectable effect 14 points). The frozen skills kept 148/150 successes on new layouts with no model and 147/150 with a DeepSeek-flash caller at USD 0.00028-0.00054 per episode; no direct-control baseline was run.

0.38

TILDE: Teacher-in-the-Loop Distillation from Embodied Foundation Models

cs.RO cs.LG ·2026-10-18 ·Yuxiang Lu, Yantong Liu, Jinghua Hou, Xueli Sun, Ruihua Han +2 more
code/data
Abstract · Project page: https://hku-sail.github.io/TILDE

Large-scale action pretraining equips Embodied Foundation Models (EFMs) with strong manipulation capabilities, with particular strengths in generalization and closed-loop robustness, but reproducing such pretraining requires substantial robot data and computation. To transfer these capabilities from pretrained EFMs to policies without action pretraining, we introduce TILDE (Teacher-In-the-Loop Distillation from EFMs), which synergizes on-policy distillation at student-induced states with learning from teacher-executed recovery trajectories. We further propose AutoDAgger, an autonomous data aggregation pipeline in which a robotic reward model is utilized to monitor task progress and trigger teacher intervention, thereby collecting recovery experience without human monitoring or manual control. AutoDAgger accelerates recovery-trajectory collection by more than $2\times$ compared with the human-in-the-loop method. By unifying learning from demonstrations, on-policy teacher supervision, and teacher's corrective behavior, TILDE improves task performance by up to $4.1\times$ over behavior cloning across three real-world tasks and supports effective cross-architecture distillation.

0.36

Generalized and Unified Power Flow Framework for Converter-Dominated AC/DC Power Grids

eess.SY ·2026-10-18 ·Hamin Chang, Fei Teng, Gyu-Sub Lee, Jae-Won Chang
code/data
Abstract · Submitted to IEEE Transactions on Power Systems. 10 pages + 10-page appendix (Appendices A-C included at the end). Code: https://github.com/…

Modern power grids increasingly rely on power electronic interfaces, requiring steady-state power flow analysis to accommodate diverse operating modes and network structures. Conventional approaches often rely on system-specific formulations and sequential approaches to handle such complexity. However, these approaches may limit applicability to diverse grid configurations and lead to convergence and computational issues. To address these limitations, this paper proposes a generalized and unified Newton-Raphson (NR) power flow framework for AC/DC hybrid grids. For generalized operational modeling, conventional and droop operating modes of generators and interlinking converters are represented through a common set of coefficients. For generalized structural modeling, models of AC transformers and DC/DC converters accommodate different transformer winding configurations and converter operating modes. For a unified solution, augmented variables are introduced to incorporate droop deadbands, output limits, and converter station losses into a single NR formulation, eliminating the need for sequential outer loops. To validate the effectiveness of the proposed framework, its results are compared with those obtained from existing power flow methods and electromagnetic transient simulations across transmission networks, distribution networks, and microgrids. The implementation is made publicly available to facilitate its use in analyzing AC/DC hybrid grids.

0.36

Scalable, Intuitive, Mixed Reality-Based Semi-Autonomous Robotic Telemanipulation for the Execution of a Fine-Grained Bimanual Task

cs.RO ·2026-10-18 ·Shifei Duan, Zhe Wang, Francesco De Pace, Minas Liarokapis
published
Abstract · 8 pages, 9 figures. Accepted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2026

Nowadays, robotic telemanipulation enables humans to control robots so as to execute tasks in a manner similar to how they perform them, and serves as an essential approach for collecting on-robot data for robot learning by human demonstrations. The quality of collected demonstrations directly impacts the performance of the learned policy. However, existing telemanipulation systems rely heavily on human operation, particularly for large-scale data collection. Collecting dozens or hundreds of demonstrations is a demanding process, as operators tend to fatigue quickly, especially during the execution of dexterous tasks that require sustained concentration. Recently, semi-autonomous robotic telemanipulation has emerged as a promising direction that reduces operator workload while preserving human-in-the-loop control for task execution. We therefore introduce a scalable, intuitive, semi-autonomous robotic telemanipulation system that leverages a mixed reality (MR) headset for on-robot data collection. The system autonomously executes the majority of the task trajectories and only requires the operator to make fine adjustments at key stages. We validate the effectiveness of our system by collecting data and training imitation learning policies on the execution of a high-precision, fine-grained bimanual peg-in-hole task. With lower data collection effort, our system obtains high-quality demonstrations that can be used to train fully autonomous task execution policies.

0.35

Energy-Score Stabilized Attentive Mixture-Density Policy: A Lightweight Approach to Multimodal Robot Motion

cs.RO ·2026-10-18 ·Sachiya Fujita, Hideyuki Ichiwara, Shigeki Sugano, Tetsuya Ogata
published
Abstract · 8 pages, 6 figures, 1 table. Accepted to IROS 2026

For practical motion generation, it is essential to handle multimodality (multiple actions can be correct for the same observation) while maintaining low inference cost and strong generalization from limited demonstrations. We introduce an attentive mixture-density policy trained with the Energy Score that represents future visuomotor trajectories as explicit probability distributions with a single forward pass. Concretely, we output Gaussian mixture parameters for joint commands (and optionally attention points) and optimize them with the Energy Score (computed as a sum of 1D CRPS under a diagonal covariance assumption) instead of the negative log-likelihood (NLL) to avoid MDN singularities and gradient-scale instabilities. Our architecture couples explicit spatial attention with mixture-density outputs, enabling robust multimodal behavior selection under visual noise and small-data regimes. On an 8-DOF manipulator, the method generates distinct avoidance trajectories and stabilizes learning in obstacle avoidance. Moreover, in a randomized moving-object grasping task, where spatiotemporal multimodality and fast timing decisions are required, the proposed model outperforms Diffusion Policy by 15-25% under several unseen conditions, demonstrating superior generalization and real-time feasibility.

0.35

Dynamic Skill-Based Routing for Call Centers with Transfers: A Value-Based Framework

eess.SY ·2026-10-18 ·Baris Ata, Ebru Kasikaralar
benchmarkcode/data
Abstract

We consider a multi-class, multi-pool queueing model of a telephone call center in which a service completion does not always end the call. The caller may be transferred to another agent for a second service, so the assignment of an agent to a call affects not only the caller's waiting time but also whether the call is resolved and how much further work it creates. We model these transfers as Markovian routing from original to rework classes, and the system manager dynamically assigns agent pools to classes so as to maximize the value generated by service, net of transfer penalties and waiting and abandonment costs, over a finite horizon. Focusing on the Halfin-Whitt heavy-traffic regime, we derive an approximating Brownian control problem. Its Hamilton-Jacobi-Bellman equation is a semilinear partial differential equation whose dimension equals the number of classes. Building on the deep splitting method of Beck et al. (2021), we develop a simulation-based computational method that approximates the value function and its gradient by neural networks, and we use it to construct an implementable routing policy for the original (prelimit) system. We calibrate the model to publicly available data from a large U.S. bank call center. In all test problems we consider, the proposed policy outperforms the best available benchmark policy we could find.

0.34

Safe Barrier Function Control for Grid-Following Inverters under Weak Grid Conditions

eess.SY ·2026-10-18 ·Grigoris Michos, Christos K. Verginis
Abstract · 10 pages

This paper proposes a barrier control approach for the safe regulation of grid-following voltage-source-converters (VSC), operating in weak grid conditions. The proposed controller aims to enhance the transient performance and synchronization properties of the VSC, providing frequency and current limiting guarantees by employing smooth control barrier functions. First, a dynamic model is derived that includes the dynamic state interactions that are omitted in quasi-steady-state approaches, revealing the destabilizing elements both during transient and steady-state performance. Then, it is analytically proven that the proposed control law guarantees frequency limitation and inner current evolution in prescribed safety regions, while counteracting the destabilizing strong coupling between PCC voltage and injected current. In addition, sufficient conditions are formulated that guarantee the stability of the closed-loop system with respect to desired operating points. The theoretic results are verified in simulations of case studies operating in varying grid strength.

0.32

GOFusion: Real-Time Gaussian-Occupancy Submap Fusion for Fisher-Informed Autonomous Robot Navigation

cs.RO ·2026-10-18 ·Boxin Bu, Simon Boche, Haoming Zhang, Jaehyung Jung, Jiaxin Wei +4 more
code/data
Abstract · 8 pages, 7 figures, 4 tables. Paper under review

In this letter, we present GOFusion, a real-time visual mapping system for autonomous robot navigation that anchors Gaussian primitives to local occupancy maps, providing both spatial awareness for path planning and high-fidelity appearance that enables information-driven robot navigation. GOFusion builds 3D Gaussian maps from pose estimates and local occupancy maps obtained via visual-inertial SLAM and enables pose corrections to maintain both geometric and visual consistency. Furthermore, it restricts the 3D Gaussian Splatting optimization to active submaps to bound GPU memory usage. Building on the appearance-rich Gaussian map, we introduce a rasterizer-integrated Fisher approximation that quantifies per-primitive sensitivity online, serving as an information-driven heuristic for Gaussian pruning and next-best-view planning. We evaluate GOFusion on public datasets and in real-world drone flights. It not only outperforms several baselines in mapping quality, but also achieves real-time, high-fidelity and uncertainty-aware mapping at scale. Finally, we demonstrate fully autonomous drone exploration for Gaussian map refinement using the proposed Fisher-informed heuristic. Code is available at https://github.com/Bennnoo/GOFusion. We provide a supplementary video demonstration at https://youtu.be/pLQKbxQvb6I.

0.32

Undulatory Swimming in Sand: Subsurface Locomotion of the Sandfish Lizard

eess.SY physics.bio-phcond-mat.softphysics.flu-dyn+1 ·2026-10-18 ·Ryan D. Maladen, Yang Ding, Chen Li, Daniel I. Goldman
published
Abstract

The desert-dwelling sandfish (Scincus scincus) moves within dry sand, a material that displays solid and fluidlike behavior. High-speed x-ray imaging shows that below the surface, the lizard no longer uses limbs for propulsion but generates thrust to overcome drag by propagating an undulatory traveling wave down the body. Although viscous hydrodynamics can predict swimming speed in fluids such as water, an equivalent theory for granular drag is not available. To predict sandfish swimming speed, we developed an empirical model by measuring granular drag force on a small cylinder oriented at different angles relative to the displacement direction and summing these forces over the animal movement profile. The agreement between model and experiment implies that the noninertial swimming occurs in a frictional fluid.

0.32

Secure Distortion-Aware GNN Precoding under Active Eavesdropping in Cell-Free Networks

eess.SP ·2026-10-18 ·Reza Ghasemi Alavicheh, Thomas Feys, Md Arifur Rahman, François Rottenberg
published
Abstract · Accepted at IEEE GLOBECOM 2026

This paper considers a cell-free (CF) network under active eavesdropping. We show that nonlinear power amplifier (PA) distortion, together with the per-access point (AP) large-scale-gain fingerprint of the attack, can be exploited across cooperating APs to enhance physical layer security (PLS). Classical zero-forcing (ZF)-artificial noise (AN) precoders rely on a linear transmitter and an unbiased estimate, and thus cannot exploit either. We propose a distortion aware graph neural network (GNN) precoder that designs the signal beams at cooperating APs and exploits both the PA distortion and the per-AP fingerprint to enhance the secrecy rate, combined with a null-space AN block to form the GNN-AN scheme. The GNN is trained on channel estimates biased by an uplink pilot-spoofing attack to maximise the average per-user secrecy rate. Simulations across different numbers of APs show uniformly higher outage secrecy rates than ZF-AN under attack, reaching gains of 15% at the 1% outage level and 8.9% at the 5% outage level at L = 32 APs, input back-off (IBO) = -1 dB, and attacker-to-targeted user pilot power ratio of 0.5, under the same signal/AN power split.

0.32

CusADi2: Latency-Optimized Batch Computation of Robot Kinematics, Jacobians, and Hessians on the GPU

cs.RO ·2026-10-18 ·John Viljoen, Lasse Peters, Masayoshi Tomizuka, Negar Mehr
benchmarkcode/data
Abstract · 8 pages, 4 figures. Submitted to IEEE Robotics and Automation Letters (RA-L). Code: https://github.com/johnviljoen/cusadi2

Trajectory optimization methods, from sampling-based zero-order to gradient-based first- and second-order approaches, require many evaluations of dynamical system functions, i.e., functions derived from a robot's dynamical model, such as forward kinematics and their Jacobians and Hessians. These evaluations are largely independent, e.g., across sampled trajectories or time steps, and can thus be evaluated in parallel as a batch. When used to make decisions online, these methods require low latency, i.e., each batch must complete in as little wall-clock time as possible. Existing frameworks utilize modern GPU hardware primarily to parallelize across batch instances, thereby optimizing throughput but not latency. We present CusADi2, a software library that reduces the batch evaluation latency by additionally parallelizing computation within each batch instance. We benchmark forward kinematics, Jacobians, and Hessians for all 67~robots in the MuJoCo Menagerie, including arms, hands, quadrupeds, humanoids, and biomechanical models, against MuJoCo XLA, MuJoCo Warp, frax, CusADi, and CasADi. CusADi2 achieves lower latency than all baselines at batch sizes of 64 to 4096, while matching them at larger batch sizes. We also demonstrate the performance advantage of CusADi2 in a sampling-based control application. All code is open-source at: https://github.com/johnviljoen/cusadi2

0.28

Low-Dimensional Hermite-Gaussian Representations for Ptychographic Reconstruction

eess.SP ·2026-10-18 ·Arya Bangun, Evan Laksono
theory
Abstract · 18 pages, 6 figures

We analyze the reconstruction error of a low-dimensional Hermite--Gaussian (HG) ptychographic method used in real-time Wigner distribution deconvolution (WDD) reconstruction. We derive an upper bound on the discrepancy between the low-dimensional, aberration-corrected reconstruction and classical Fourier-domain WDD. The bound decomposes the reconstruction error into contributions from HG--Fourier basis mismatch and aberration-coefficient mismatch, explicitly characterizing their dependence on the number of retained HG modes, Wiener regularization, aberration mismatch, and measurement dose. Numerical experiments on simulated electron and X-ray ptychographic data characterize these dependencies and demonstrate robustness to Poisson noise. These results provide theoretical and guidance for selecting the reduced representation and reconstruction parameters, enabling computationally and memory-efficient for near-real-time ptychographic reconstruction.

0.26

Stability and Stabilization of Kronecker-Sum-Coupled Multilinear Systems

eess.SY math.OC ·2026-10-18 ·Ziqin He, Ruoying Wang, Shenghan Mei, Yidan Mei, Xin Mao +1 more
unusual
Abstract · 6 pages, 4 figures

Coupled dynamical systems with multi-dimensional state structures arise in many applications such as spatiotemporal processes and large-scale interconnected systems. However, existing studies primarily focus on matrix-based coupled systems or tensor-based dynamical systems without couplings. The properties of coupled tensor-based systems remain largely unexplored. In this article, we study Kronecker-sum-coupled multilinear systems under the transform-based tensor product framework. We establish fundamental tensor spectral properties showing that stability can be characterized through lower-dimensional subsystem dynamics together with slice-wise matrix stability in the transform domain. We further develop optimization conditions for state-feedback stabilization and show that the stabilization problem can be decomposed into independent subsystem problems. Finally, we analyze the computational complexity of the proposed framework and demonstrate its substantial computational advantages over unfolding-based approaches using both synthetic systems and multi-inertia motor systems.

0.25

AI-Enhanced Radiation Classification in ROS

cs.RO ·2026-10-18 ·Youngseo Lee, Cesar Alan Contreras, Tianshu Ruan, Manolis Chiou, Rustam Stolkin
Abstract · 7 pages, 4 figures, 3 tables

This study presents the development of two Robot Operating System (ROS) nodes with a primary focus on differentiating seven radioactive isotopes. The first node, referred to as the model training node, uses a Gradient Boosting classifier to identify and classify these isotopes. The classifier achieved a reported overall accuracy of 95\% on the held-out test dataset. The second node, known as the prediction node, provides classification results when users input data in CSV format. For the practical demonstration, the C12137 radiation detector module was mounted on the Jackal UGV, and an operator approached different radiation sources to collect data. To support future dataset expansion, we also developed a method for converting gamma-spectrum graph images into CSV files. In a separate robotic evaluation across three source locations, 30 of 42 trials were correctly classified (71.43\%). Am-241, Co-60, Cs-137, Na-22, and Ti-44 were correctly classified in all six trials each, whereas Ba-133 and Eu-152 were incorrectly classified in all six trials each. These results show lower accuracy in robotic trials than in the held-out test set.

0.25

Receiver Design for Continuous-Time AFDM based on Pre-Filter De-Chirp

eess.SP ·2026-10-18 ·Qingyu Li, Shuangyang Li, Guanghui Liu, Fuchen Xu, Hongjun Liu
hot direction
Abstract · Submitted to IEEE International Conference on Communications (ICC 2027)

Affine frequency division multiplexing (AFDM) is a promising waveform for future high-mobility communications due to its Doppler robustness and diversity gain. However, existing continuous-time AFDM relies on chirp-dependent wideband receive filtering, which increases noise power and causes channel-energy leakage. In this paper, we propose an AFDM reception scheme based on pre-filter de-chirp for doubly selective channels. The coupling between de-chirp and the channel is first characterized, yielding an equivalent path representation with delay-coupled effective Doppler shifts. The discrete affine Fourier transform (DAFT)-domain equivalent channel is then derived under general transmit pulse shaping and linear time-invariant receive filtering, showing that the effective filtering kernel depends only on the relative sample offset and avoids the additional spreading caused by conventional chirp-dependent filtering. The receive-noise power is further analyzed, and an explicit expression for the signal-to-noise ratio (SNR) gap is obtained by the required receive-filter bandwidths. Simulations verify the theoretical SNR analysis and show that the proposed scheme significantly outperforms the conventional scheme while approaching ideal discrete-time AFDM.

0.25

Structure over Reasoning: Unified Zero-Shot Navigation with a Small Vision-Language Model

cs.RO ·2026-10-18 ·Zeqi Li, Xunyi Zhao, Zerui Li, Sihao Lin, Qi Wu +1 more
Abstract · 8 pages, 1 figure

Vision-language navigation is moving into photorealistic, physically simulated scenes, where the leading zero-shot agents put a vision-language model in the control loop to choose every move and stop. Robots favour small models, whose abilities are uneven: Qwen3-VL-4B can judge whether a single view looks like the goal, but its route reasoning is at chance, its distance estimates nearly constant, its stop votes unreliable. Our agent therefore asks it two perceptual questions per step, which direction looks like the goal and whether the view justifies stopping, plus one gated frontier query on route-bearing instructions, and treats the answers as evidence, not actions: a route-blind semantic field over a bird's-eye map accumulates the directional scores through a tempered update, a skeleton graph supplies frontiers and geodesic distances, and a two-stage terminator turns stop votes into termini through a trail-consistency gate and a budget-forced return. With every baseline re-run in one installation at one budget, the agent exceeds all four zero-shot baselines of the VLNVerse subset in mean success on both granularities, though the paired interval against the strongest coarse-grained baseline includes zero. Termination is the dominant lever in ablation, thirteen attempts to give the model more of the navigation establish no reliable gain, and the agent runs on a Unitree Go2 quadruped with off-board inference. Code, per-episode logs and a robot video will be released at https://anonymous.4open.science/r/structure-over-reasoning-PAGE.

0.24

Intent-Informed Model Predictive Control for Shared Autonomy in Remote Manipulation Tasks

cs.RO ·2026-10-18 ·Lorena Maria Genua, Camilla Tabasso, Mya Schroder, Joshua Migdal, Tye Brady +1 more
Abstract

Shared autonomy is a control paradigm that combines human cognitive abilities with robot precision to enable effective remote manipulation in complex environments. A key challenge is how to adapt the level of autonomy using information extracted from the human, task, and/or environment. In this paper, we propose a novel arbitration strategy that dynamically adjusts the level of autonomy in the system using the entropy of a belief distribution derived from a multimodal intent inference. Human intent is modeled with a Dynamic Bayesian Network (DBN), which updates beliefs over goals given real-time operator gaze data and robot arm motion during teleoperation. We then formulate the blending of human input and autonomous control as a Model Predictive Control (MPC) problem, where the system's autonomous behavior is guided by the belief distribution. This approach enables continuous, gradual transitions across the full spectrum of autonomy levels, from direct teleoperation to full autonomy. To validate our approach, we conducted experiments to compare the performance of our proposed algorithm against two MPC-based shared autonomy methods. Compared to the baselines, our approach achieved reductions of at least 61% in object retrieval time, 44% in trajectory length, and 60% in human control input. Participants reported high perceived control, predictability, intent recognition, and responsiveness, along with low workload.

0.24

Normalizing Flow for Neural-Enhanced Track-Before-Detect

eess.SP cs.LG ·2026-10-18 ·Shaoxiu Wei, Bhaskar Rao, Florian Meyer
theory
Abstract

Multiobject tracking (MOT) is a fundamental capability in autonomous navigation, radar and sonar systems, as well as 6G integrated sensing-and-communication (ISAC) networks. In this paper, we propose a neural-enhanced track-before-detect (TkBD) framework that operates directly on complex-valued raw signals and learns a latent representation of unmodeled propagation effects during runtime. We introduce an environment-dependent latent intensity variable that characterizes the signal power of each object in the shared propagation environment. The probability density function of this intensity variable is represented by a normalizing flow and trained online in an unsupervised manner within the sequential Bayesian inference procedure. The learning objective is derived from a variational Bayes formulation by maximizing the evidence lower bound (ELBO), thereby accounting for the uncertainties inherent in the Bayesian TkBD problem. To enable scalable inference, we make use of a recently introduced Bayesian TkBD framework that combines belief propagation, which avoids high-dimensional operations, with sequential Monte Carlo methods for efficient low-dimensional updates.

0.23

ChronoChart: Setup-Agnostic Channel Charting for Label-Free Indoor Localization

eess.SP cs.LG ·2026-10-18 ·Majid Sharghi Foroushani, Jonas Pirkl, Jonathan Ott, Maximilian Stahlke, George Yammine +2 more
published
Abstract · To be published in the proceedings of the 16th International Conference on Indoor Positioning and Indoor Navigation (IPIN)

Accurate radio-based indoor localization remains challenging in multipath-rich and dynamic indoor environments. Supervised fingerprinting methods achieve high positioning accuracy by learning mappings from radio channel measurements to physical locations, but require costly position labels and frequent recalibration. Channel Charting (CC) addresses this limitation by learning low-dimensional representations of Channel State Information (CSI) without position labels, yielding channel charts in which nearby points correspond to nearby physical locations. However, prominent CC methods rely on setup-specific dissimilarity metrics, such as ADP-, CIRA-, or PDR-based distances, and on precomputed $O(n^2)$ distance matrices, limiting scalability and transferability across radio systems. We propose ChronoChart, a timestamp-supervised CC framework that requires only CSI and timestamps without prior knowledge for training. Our method jointly optimizes complementary objectives: a triplet loss using temporally derived positive and negative samples to establish coarse global structure, a temporal-proximity loss that promotes smooth trajectories in the chart, and a joint-embedding loss that enforces consistency between representations of different augmentations. Evaluated on real-world 5G, MIMO, and passive UWB data, our method matches or exceeds prior CC baselines that depend on tailored distance metrics, and remains scalable and applicable across heterogeneous radio systems.

0.22

Self-Crystallizing Behavior Trees: Amortizing Foundation-Model Reasoning into Deterministic Control for Autonomous Drones

cs.RO ·2026-10-18 ·Behnam Moradi
Abstract

Classical robot control, Finite State Machines and Behavior Trees (BTs), is deterministic and inspectable but fixed at design time: it fails on situations its designers did not anticipate. Foundation models can synthesize new control logic at runtime, at a cost in latency and non-determinism hard to accept in safety-critical aerial systems. Existing LLM-driven approaches keep the model in the loop permanently and discard what it produces. We propose Self-Crystallizing Behavior Trees (SCBT), in which the foundation model is a compiler rather than a runtime. When a BT node fails, an LLM proposes a minimal, mission-agnostic subtree patch; CMA-ES tunes its parameters in simulation; a contextual bandit decides whether a stored patch or a new LLM call is worth its cost; and a three-stage admission pipeline (static check, invariant-monitored replays, an empirical margin over the unpatched tree) admits only patches that demonstrably recover. Admitted patches are stored in a fleet library and, once the LLM call rate decays, crystallized into a deterministic BT that no human wrote. We implement SCBT on Microsoft Project AirSim with a schema-constrained LLM pipeline (Gemini 3.1 Pro) and a lockstep simulator. Over three seeds of a 20-episode learning phase with wind-gust and battery failures, SCBT used $11\pm1.4$ LLM calls against 26 for whole-tree regeneration and 22 without a library, with 0.33 calls per episode in the second half. The crystallized trees, 28 and 33 nodes with two patches each, reached $97\pm2\%$ mission success with no learning running, against 12\% for the unpatched tree and 98\% for a hand-coded expert tree, and reproduce decision traces exactly when re-ticked offline. Two negative results: replay admission rejected useful wind recoveries by attributing wind-induced geofence excursions to the patch, and a static-only ablation generalized better to held-out failures.

0.22

Dream What I Can: Sparse Visual Planning for Long-Horizon Robot Manipulation

cs.RO ·2026-10-18 ·Chongkai Gao, Wenlong Huang, Haonan Chen, Lin Shao, Yilun Du +3 more
Abstract

Video models provide rich priors for generalizable robotic manipulation, but plausible visual futures are not always executable. We introduce DreamCan, a hierarchical policy that grounds video models through sparse visual planning: a high-level planner adapted from a pre-trained video model and a low-level visuomotor policy. Rather than predicting dense trajectories, the high-level planner generates sparse visual subgoals that balance reachability and task utility. From offline, action-free robot videos, we estimate these quantities and use dynamic programming to select subgoal sequences without manual annotations. The resulting subgoals are more reliably attained than uniformly spaced alternatives, and the full system outperforms language, dense-video, and uniform-subgoal planners, reaching high planning success rates on Language-Table tasks and 91.2% success on LIBERO-Long with a 0.1B-parameter low-level policy. Additional suboptimal videos and test-time compute further improve success. The system also performs challenging real-world bimanual tasks, including box folding and packing with unseen object combinations. More results are available at https://dreamcan.github.io/.

0.22

Finite-Time Capture by a Dubins Pursuer Against Unknown Evader Strategies

eess.SY cs.RO ·2026-10-18 ·Dinesh Patra, Bhargav Jha, Ashish R. Hota
Abstract

We consider a planar pursuit-evasion problem in which a Dubins pursuer seeks to capture a slower first-order point-mass evader that can change its heading instantaneously. First, we propose a turn-rate feedback law with a single tunable gain and derive an explicit admissible gain interval that guarantees finite-time capture against any admissible evader strategy while satisfying the turn-rate constraint. Next, we obtain an explicit capture-time bound and characterize the turning authority required for the gain interval to be non-empty. Finally, we show that the dependence on the evader's speed bound can be removed by deriving a speed ceiling, computable solely from the pursuer's parameters, below which the guarantees hold at the maximum allowable gain. Numerical simulations illustrate the effectiveness of the proposed approach.

0.20

The Bitter Lesson Comes to Robotics

cs.RO ·2026-10-18 ·Christopher Leet, Achyutha Menon, Sravanthi Machcha, Sven Koenig, Tzu Kit Chan +1 more
benchmark
Abstract · Submitted to the Workshop on The Action Gap (TAG) at CoRL 2026, updated with Gemini Flash 3.6 and Gemini 4 Argon

General purpose multimodal language models (LMs) are increasingly capable of directly controlling robotic arms (e.g., by commanding end effector poses) to manipulate their environment. Tracking the development of these capabilities is important for understanding their potential benefits, such as improving industrial automation, and risks, such as enabling malicious users to produce dangerous chemical or biological agents. To this end, we investigate the ability of frontier multimodal LMs to directly control physical robotic arms. We evaluate nine models at multiple reasoning-effort and service-tier settings on six manipulation tasks taken from the RoboDojo benchmark. To our knowledge, this is the first systematic comparison of a broad range of frontier LMs across a common suite of manipulation tasks on physical hardware. We find that current frontier models can directly control real robotic arms to perform a variety of manipulation tasks but are not yet precise or efficient enough to complete them within a practical action budget. However, these capabilities are improving rapidly: performance increases on our benchmark between July and October 2026 (inclusive) correspond to a performance doubling time of 53.9 days, while the best recorded execution rate increased by a factor of 4.4 and the lowest recorded execution cost fell by a factor of 6.2. These trends raise the possibility that the Bitter Lesson is coming to robotics; the capabilities of embodied general purpose LMs may one day be comparable to those of more specialized models.

0.19

AeroDASH: Asynchronous Vision-Language-Action Control for Continuous UAV Flight

cs.RO ·2026-10-18 ·Zhe Liu, Yuxuan Zheng, Jinzhe Qin, Ziqi Gao, Boyu Zhou
Abstract · 8 pages, 6 figures, 2 tables

Onboard inference latency creates a mismatch between the trajectories predicted by vision-language-action (VLA) policies and the ongoing motion of unmanned aerial vehicles (UAVs). We present AeroDASH, an asynchronous VLA framework for continuous language-conditioned flight. A B-spline policy is initialized through supervised fine-tuning on expert demonstrations, providing continuous trajectory references in position and yaw. During execution, trajectory fusion removes the expired prediction prefix and connects the remaining segment to the active reference with derivative continuity while limiting trajectory deformation. To account for deviations induced by asynchronous execution, execution-aware group relative policy optimization (GRPO) further adapts the policy using trajectory-level rewards. On the UAV-Flow test set, AeroDASH achieves a success rate of 76.56%, exceeding OpenVLA-UAV and AeroDASH w/o GRPO by 38.1 and 20.1 percentage points, respectively. Execution experiments show that AeroDASH sustains motion during inference and reduces acceleration, jerk, and speed fluctuations relative to asynchronous tracking of waypoint sequences. We further demonstrate language-conditioned flight with onboard inference and asynchronous trajectory execution. The code will be released upon acceptance.

0.17

Distributionally Robust Chance-Constrained Reinforcement Learning for Home Energy Management under Electric Vehicle Uncertainty

eess.SY ·2026-10-18 ·Meng Yuan, Ye Wang, Yi Mei, Ramesh Rayudu, Alan Brent +2 more
Abstract

The electrification of residential energy systems and adoption of electric vehicles (EVs) create demand-side flexibility. However, random driving behaviour makes it difficult to ensure EV readiness at departure and maintain thermal comfort. To address this, we formulate residential scheduling as a distributionally robust (DR) chance-constrained Markov decision process, whose 2-Wasserstein ambiguity set guards against tail risk and distribution shift in EV behaviour. The intractable DR chance constraint is then replaced by a sufficient tightened expectation constraint under the nominal distribution. The resulting problem is solved by a Lagrangian soft actor--critic with distributional cost critics, and an online boundary guard enforces EV reachability and thermal comfort within the actuator limits at every step. Five controllers are compared over one year in Wellington, New Zealand, under a degraded forecast. The proposed controller is the cheapest controller that meets the EV target on every day and both chance constraints under forecast error and random driving behaviour. Compared with the rule-based baseline, the only other controller to do so, it lowers the annual cost by 8.6% and the comfort violation hours by 69%. The ablation study confirms that the boundary guard ensures pointwise feasibility within the actuator limits, and the DR chance constraint reduces tail risk.

0.16

Visual-Inertial Odometry for Ground Robots with Dynamic Camera-IMU Extrinsic Calibration

cs.RO ·2026-10-18 ·Bo Dong
Abstract

Visual-inertial odometry systems are often constrained by the assumption that sensors are rigidly mounted, which limits their adaptability in practical scenarios involving structural motion. To address this limitation, this paper investigates time-varying camera-IMU extrinsic calibration. First, a dynamic camera-IMU extrinsic calibration method based on the Denavit-Hartenberg convention is proposed by incorporating the robot kinematic model. Second, a two-step initialization strategy is developed to efficiently estimate the initial joint angles, providing a reliable starting point for subsequent extrinsic parameter optimization and enabling online estimation of timevarying camera-IMU extrinsic parameters. Simulation experiments demonstrate that the proposed framework can effectively accommodate variations in camera-IMU extrinsic parameters while maintaining competitive localization accuracy.

0.14

Localization and Neural Splatting on the Edge: An FPGA Scene Memory for Robots

cs.RO cs.CV ·2026-10-18 ·Haohua Que, Mingkai Liu, Tianle Zhu, Handong Yao
Abstract · 8 pages

Resource-constrained robots and autonomous vehicles need scene memories that localize a camera and render nearby views without a desktop GPU. Scene-coordinate regression (SCR) and 3D Gaussian splatting (3DGS) fit because they avoid an SfM stage, but ACEsplat-style pipelines are GPU-centered, and moving them to an edge FPGA is hard because SCR is pose-sensitive and 3DGS rendering is irregular and capacity-sensitive. We present LENS, an edge scene-memory acceleration system on the AMD KR260 that brings an ACEsplat-style workload to a low-power FPGA/ARM platform. Rather than proposing a new scene model, LENS co-designs the two dominant kernels with the board data path: an INT16 SCR accelerator that preserves pose accuracy, and a bit-exact fixed-point ordered 3DGS renderer that rewrites alpha transmittance as a log-sum-exp recurrence and removes provably useless streamed Gaussians. All LENS latency, power, and quality results are board-measured. The SCR engine preserves DSAC* localization accuracy at 25.42 ms, the on-board renderer reaches 34.11 dB on Cambridge against ACEsplat's 33.20, novel-view synthesis on RealEstate10K reaches 27.84 dB against 26.11, and all ten Wayspots scenes reach the ACEsplat rendering target. LENS wins performance per watt over a Jetson Orin NX on every deployment task, by 1.9x to 5.7x at about a quarter of the power, and is faster in absolute latency on localization and Wayspots rendering. On-device training reaches ACEsplat-level quality but not the GPU's time or energy, marking the remaining boundary for a future training accelerator.

0.14

Adaptive Bendpoint Formation in Octopus-Inspired Soft Continuum Arms

cs.RO ·2026-10-18 ·Jessica Gumowski, Julian Finn, Krishna Manaswi Digumarti, David Howard
Abstract

Octopuses exhibit a powerful form of embodied intelligence from a highly adaptable soft body. One key mechanism is the ability to instantiate bending points at any location along the arm, which enables performant and dexterous behaviours. Taking inspiration from the octopus for its dexterity, this study focuses on changing the curvature of a tendon-actuated arm by adaptively changing the anchoring end-point along the arm. An array of inflatable shells is integrated into the arm, where inflating a shell causes an increase in friction at the location, jamming the tendon and creating a temporary new anchoring point. The novel soft actuator does not break the continuity of the arm between the actuated segment and the loose segment. Experiments show that the actuated segment has a non-constant bell-shaped curvature which is comparable to that of a real octopus arm during a reaching motion. Compared to conventional single-tendon continuum manipulators, the proposed design offers an expanded workspace and enhanced dexterity through the ability to independently shape different regions of the arm through selective curvature tuning. At the same time, it avoids the mechanical complexity typically associated with multi-segment, multi-actuator systems.

0.14

DexGraspFlow: Topology-Aware SE(3) Flow Matching for Cross-Embodiment Dexterous Grasp Generation

cs.RO ·2026-10-18 ·Bin Zhou, Boyu Ma, Xinyu Zhou, Jiahua Dong, Chuhao Zhou +3 more
Abstract

Differences in hand geometry, joint parameterization, and kinematic topology pose challenges for cross-embodiment dexterous grasp generation. However, spatial relations among links alone are insufficient to fully capture the hand's kinematic topology. We propose DexGraspFlow, a conditional link-level $\mathrm{SE}(3)$ flow-matching framework integrating geometric relations and kinematic topology. The framework encodes local object geometry, link geometry, and current poses. A dual-stream DiT models link--link coordination and link--object interaction, while attention biases derived from kinematic-tree distances provide explicit structural information for modeling evolving geometric interactions. The model jointly learns translational and rotational velocity fields, integrates them to generate link poses, and recovers joint configurations through inverse kinematics. Additionally, inference-time palm-orientation guidance enables controllable grasp generation without retraining. Across three robotic hands and 30 object--hand pairs, DexGraspFlow achieves a 95.33\% average simulation success rate, with an average generation latency, including inverse kinematics, approximately 49\% lower than that of a state-of-the-art baseline. Further experiments demonstrate zero-shot adaptability to seven finger-removal variants and palm-orientation control. A model trained on demonstrations for the deployed hand achieves a 91\% success rate over 100 real-robot grasping trials, supporting the framework's applicability to physical grasp execution.

0.14

T(R,O) Grasp++: Cross-Embodiment Dexterous Grasping in Cluttered Environments at Scale

cs.RO ·2026-10-18 ·Xin Fei, Zhixuan Xu, Huaicong Fang, Fangjun Chen, Zhaoyu Mu +3 more
Abstract

Dexterous grasping in cluttered scenes remains a central challenge in robotic manipulation, as it requires high-dimensional hand coordination under complex object-environment constraints. To address this problem, we introduce one of the largest simulation benchmarks to date, comprising around 0.5B grasp annotations across eight dexterous hands and approximately 50K objects. Furthermore, we propose T(R,O) Grasp++, a scalable diffusion-based framework for cross-embodiment dexterous grasp synthesis in cluttered environments. Extensive experiments demonstrate that our model significantly outperforms existing baselines, achieving an average success rate of 88% across eight dexterous embodiments. Meanwhile, our pipeline achieves only 0.3s inference latency while improving inference sample efficiency by over 10 times.

0.13

Learning Waypoints and Passage Events for Decentralized Multi-Quadrotor Trajectory Planning

cs.RO eess.SY ·2026-10-18 ·Yunhan Shi, Jia Deng, Zehao Jia, Lin Cheng, Shengping Gong
Abstract · 9 pages, 7 figures, 3 tables

Multi-quadrotor passage planning requires coordinated spatial routes and passage times under vehicle dynamics and actuation constraints. Optimizing these decisions jointly couples temporal allocation, nonlinear motion, and inter-vehicle interactions, creating a computational challenge for local trajectory generation. We propose a learning-optimization framework that predicts spatial waypoints and ordered passage events for independently solved quadrotor planners. A common physical time grid aligns the vehicles, with intermediate events fixed by the networks and terminal progress optimized locally to permit different completion metrics. Differentiable sequential convex programming incorporates rigid-body dynamics and rotor-thrust bounds, allowing collision and completion gradients to train both networks through the generated trajectories. In a two-quadrotor evaluation across 24 layouts, the adapted policy passes physical feasibility and sampled separation checks in all five held-out layouts. Their mean completion metric decreases by 1.52% relative to the inherited policy. The median forward time is 21.01 ms on the held-out exchange layout with matching numerical trajectory references. These results demonstrate that learned spatial and event parameters can coordinate independent rigid-body planners while preserving local vehicle constraints in the evaluated scenarios.

0.13

Learning When to Adapt: Reactivity-Conditioned Residual Adaptation under Human-Robot Distribution Shift

cs.RO ·2026-10-18 ·Immanuel R. Santjoko, Benedictus C. G. Cinun, Jing Chen, Bin Hu
Abstract

Pedestrians differ in whether they respond to a robot, yet many RL policies assume all are non-reactive to prioritize safety. Policies trained without conditioning on local pedestrians' reactivity may therefore be unnecessarily conservative or rely on cooperation that does not occur. We propose RCRA, an online adaptation framework that modulates a learned residual around a frozen conservative navigation policy according to estimated pedestrian reactivity. A Transformer-based estimator infers latent pedestrian reactivity from motion histories and robot context, providing a more accurate signal for navigation adaptation and improving scenario-averaged simulation accuracy by $12.62$ percentage points over the Bayesian baseline, which assumes known pedestrian goals. These estimates are spatially aggregated into an adaptation scale that gates LoRA modules, enabling cooperative behavior near reactive pedestrians and conservative behavior near non-reactive pedestrians. Among baselines achieving SR above $90\%$ in every scenario, RCRA reduces scenario-averaged path length by $16.7\%$ relative to the conservative backbone and by $7.3\%$ relative to randomized-training baselines. It improves mean SR by $14.88$ percentage points over the reactive-trained policy.

0.11

AQUAACT: A Joystick-Level Underwater Teleoperation Dataset and Closed-Loop Benchmark

cs.RO ·2026-10-18 ·Zhenqi Wu, Trung Tien Dong, Aditya Penumarti, Mohamed A. Fouad, Menna A. Fouad +3 more
benchmark
Abstract

Underwater robots support a growing range of work, from inspecting offshore structures and pipelines to surveying reefs and accompanying divers. Most of these tasks, however, are still flown by hand by a pilot, which raises the question of whether they can be carried out autonomously instead. On land, vision-language-action (VLA) models already let robots carry out tasks autonomously from language instructions, which makes them a natural candidate for taking the pilot's place. Underwater, however, there is little data to train such models. We take inspiration from how these tasks are flown today: the pilot watches the camera image and steers with a joystick, whose commands the flight controller turns into motion. We therefore record the pilots and introduce AQUAACT, to our knowledge the first publicly available language-annotated underwater dataset pairing camera images with human joystick commands. It comprises 1.9 hours of real-world and 17.9 hours of simulated demonstrations, both flown by human pilots, together with the vehicle's inertial measurements. Experiments with AQUAACT show that its simulated demonstrations transfer to real data and greatly reduce the real demonstrations a policy needs. A closed-loop benchmark of seven policies in simulation, from imitation-learning baselines to VLA models, each trained with its default recipe, shows that these tasks remain far from solved: no policy completes more than a fifth of the episodes. Pretraining on land data gives large VLA models no advantage over small policies trained from scratch, a gap that land-trained models have yet to close underwater.

0.07

SurGPT: A Multi-Task Foundation Model for Robotic Surgical Skills Assessment and Gesture Recognition

cs.RO ·2026-10-18 ·Parham M. Kebria, Saeid Nahavandi, Mehrdad Saif, Fernando Bello
Abstract

The assessment of surgical skills in robotic surgery training has traditionally relied on subjective evaluations by human experts, creating bottlenecks in surgical education and limiting scalability. We present a novel foundation model that automatically analyzes robotic surgical videos to provide comprehensive skill assessment and gesture understanding. Combining a visual encoder with a transformer architecture, the model performs three key tasks simultaneously: 1) surgical task classification (knot tying, needle passing, suturing), 2) skill level assessment on a standardized scale, and 3) natural language generation of surgical gesture descriptions. The model processes video sequences from the da Vinci surgical robot and generates human-readable text descriptions of surgical actions, similar to how language models generate text from prompts, using a multi-task learning framework with adaptive uncertainty weighting to optimize performance across all objectives. Our transformer-based architecture incorporates positional encoding for temporal understanding and attention mechanisms for spatiotemporal feature integration. Evaluated on the da Vinci Skills Simulator dataset, the proposed framework demonstrates state-of-the-art performance in surgical task recognition and provides interpretable skill assessments that correlate with expert evaluations, with the model's ability to generate natural language descriptions of surgical gestures opening new possibilities for automated surgical education, real-time feedback systems, and objective skill assessment in surgical training and certification programs.

0.06

People & Society 16

Complementary Vulnerabilities: Auditing and Governing Clinician-LLM Error Overlap

cs.CY ·2026-10-18 ·Mohsen Amoei, Kathleen Fraser, Dan Poenaru
published
Abstract · 4 pages main text, 2 figures. Accepted at SMASH 2026, the Symposium on Model Accountability, Sustainability and Healthcare (Mila, Montreal, …

Health systems are deploying large language models (LLMs) into clinical workflows, often on the assumption that a second automated check can only reduce error. But a second reasoner reduces error only insofar as it fails differently from the first. We say that a clinician and a model have complementary vulnerabilities to the extent that they break down through different mechanisms and on different cases. The two reasoners' failure substrates differ: they are embodied and cognitive in the clinician, and statistical and generative in the model - so their weak points need not line up. That divergence is plausible rather than guaranteed, and the independence it creates is fragile, eroding when clinicians defer to the model or the model conforms to the clinician. We therefore propose a predeployment audit in which each site measures error overlap on its own cases. We also propose a governance model with which the site keeps those errors independent over time. Health systems should then judge a clinical artificial intelligence (AI) tool by whether its failures overlap with their clinicians' failures, as well as by its accuracy.

0.37

It's Complicated: Factors Influencing the Effectiveness of LLM-Generated Feedback in a Live Programming Course

cs.HC ·2026-10-18 ·Tsvetomila Mihaylova, Arto Hellas, Bita Akram, Narges Norouzi, Peter Brusilovsky +1 more
top venue
Abstract

We report on three iterations of an LLM-generated feedback system deployed in a live web development course, spanning three underlying models and three feedback formats. We ask how feedback presence and its characteristics (length, complexity, format) relate to student performance, whether prior feedback exposure is associated with final-project performance, and how performance changes after first exposure. Feedback was associated with more favorable immediate outcomes in several comparisons, but the pattern varied by outcome and period. In pooled models, longer feedback was associated with lower next-submission correctness, more remaining attempts, and longer resubmission time. Prior same-part feedback was associated with lower final-project solve probability. Student-level comparisons found differences on final-project steps begun before exposure, while within-student changes across steps were uncertain. These comparisons do not identify the causal effect of feedback. Changes on standard exercises were more favorable among exposed students, but the groups' different course-stage splits limit causal interpretation. Bounded message length and student choice of format are directions for experimental evaluation. Our final-project analysis captures follow-up performance rather than directly measuring learning or understanding.

0.36

Voice-Based AI Feedback in Higher Education: A Field Study on Quality and Actionability

cs.CY cs.HC ·2026-10-18 ·Thilo Tamme, Leander Schilling, Anton Hantel
published
Abstract · 9 pages. Research proposal accepted, following double-blind peer review, for presentation at the 46th Annual Conference of the Strategic Man…

Student feedback is a critical input for continuous improvement in higher education, yet traditional end-of-semester written surveys suffer from low response rates, superficial answers, satisficing, and limited actionability. Voice-based conversational AI agents can conduct structured interviews at scale, asking open-ended questions and probing follow-ups, and may capture richer qualitative feedback than written instruments. This research proposal asks what characteristics, depth, and strategic value feedback collected through voice-based conversational AI agents has in higher education, and whether such agents elicit detailed, honest, and actionable feedback that can support instructional improvement and institutional learning. Drawing on survey-mode-effects research, the Computers As Social Actors (CASA) paradigm, and dynamic capabilities theory, we develop three hypotheses on elaboration, social presence and disclosure, and actionability. A completed technical pilot at the Technical University of Munich established the feasibility of telephone-based AI interviews; the main field study is planned for the "Advanced Leadership Communication" course at MIT Sloan with approximately 60 students, using a convergent mixed-methods design and a four-dimensional coding framework for actionability (specificity, thematic richness, constructive orientation, contextual depth). The study extends CASA theory to evaluative educational settings, characterizes voice-based AI as a distinct feedback modality, and links micro-level feedback quality to institutional sensing capabilities.

0.34

When Is Disclosing Private Conversations with AI to the State Justified? Proactive Referral Policies of Major U.S. AI Companies through the Lens of Contextual Integrity

cs.CY ·2026-10-18 ·Minjong Cheon, Hanseon Joo, Ook Lee, Jongchang Ahn
Abstract · 38 pages (22-page main text followed by 16-page supplementary material). Minjong Cheon and Hanseon Joo contributed equally

AI companies engage in "proactive referral," notifying law enforcement of user conversations they judge dangerous without a government request. This poses a dilemma, since referral may expose people who merely confided dangerous thoughts to an AI to investigation, whereas not referring may leave actual violence unprevented. This study asks under what conditions proactive referral can be justified and to what extent companies' public policies disclose those conditions. Drawing on contextual integrity theory and the norms governing exceptional information sharing in confidential counseling, online posts, and private communications, the study develops eight justification criteria classified as required, presumptive, or recommended. Three evaluators analyzed 27 public documents from six major U.S. AI companies using the criteria applicable to corporate policies. No company clarified whether broad sharing provisions or narrower safety notices took precedence, and assessments of referral scope depended on which provisions were given precedence. No company disclosed the specific factual thresholds and exception conditions governing referral decisions, or whether users would be notified afterward. The study turns contextual integrity into a stage-by-stage framework for proactive referral, shows that policy documents must be read in light of their relationships, and proposes policy elements for keeping proactive referral a narrow exception disclosed in advance.

0.34

On the near-field evolution of a chemically reactive plume behind a ship

cs.CE ·2026-10-18 ·Ramesh Kolluru, Savvas Gkantonas, B Harikrishnan, Epaminondas Mastorakos
unusual
Abstract · 15 pages, 7 figures

The near-field evolution of chemically reactive pollutants in the exhaust plume of a 270 m crude oil tanker is investigated using a high-fidelity Computational Fluid Dynamics (CFD) framework coupled with a Conditional Moment Closure (CMC) turbulence-chemistry interaction model and the CBM-IV atmospheric chemistry scheme. Two complementary approaches, a canonical zero-dimensional CMC (0D-CMC) and a spatially resolved Imperfectly Stirred Reactor Network (ISRN), quantify micromixing effects using information from the CFD flow field. Micromixing strongly modulates the evolution of NO, NO2, O3, and PAN, explaining spatial discrepancies between detailed CFD-CMC results and Gaussian plume estimates. While conventional Gaussian formulations overpredict near-field NO by up to 45% within the first 300 m, they are used here as a contextual reference to highlight where turbulence-chemistry coupling matters. The CFD-CMC-ISRN framework therefore provides a physically consistent basis for integrating micromixing effects into next-generation dispersion and air-quality models. This framework bridges the gap between combustion-derived turbulence-chemistry models and environmental dispersion simulations, enabling improved parameterization of subgrid mixing in regulatory contexts.

0.32

A $u$-$p_w$ SPH framework with mechanical-hydraulic boundary treatment in GeoDualSPHysics for large-deformation geomechanics

cs.CE ·2026-10-18 ·Yangfan Ma, Bin Su, Zheng Han, Changze Li, Mitsuteru Asai
Abstract · 23 pages, 11 figures, 6 tables

Pore-water pressure changes the effective stress and deformation of saturated geomaterials, while skeleton deformation generates pressure and drives drainage. Resolving this feedback together with hydraulic boundary conditions is necessary for simulations that connect consolidation to large post-failure movement. This paper presents a hydromechanical extension of GeoDualSPHysics using a single-layer displacement-pore-water-pressure ($u$-$p_w$) smoothed particle hydrodynamics (SPH) framework. An explicit pressure-rate equation combines volumetric compression and Darcy diffusion, and the resulting pore pressure feeds back into the skeleton momentum balance. The implementation extends the existing modified dynamic boundary condition (mDBC) infrastructure to carry hydraulic ghost values alongside mechanical boundary data. Prescribed pressure at drained surfaces is combined with impermeable walls and impermeable loaded contacts, with separate control of loading and drainage activation. The formulation is assessed through one-dimensional consolidation, two-dimensional flexible-strip loading, and the three-dimensional Cryer problem. The Cryer simulations reproduce the non-monotonic central pressure response over four Poisson ratios, with normalized pressure root-mean-square errors of 0.012-0.030. The consolidation tests also expose remaining settlement and late-time pressure discrepancies. Applications to 5 m and 8 m sensitive-clay slopes illustrate sequential retrogressive failure and large deformation. The contribution is the integration of hydraulic feedback and boundary handling within GeoDualSPHysics; the results establish an initial verification basis while identifying remaining accuracy and convergence questions.

0.31

AI-Mediated Employee Voice and Performance: Disentangling Informational and Affective Pathways

cs.CY cs.HC ·2026-10-18 ·Thilo Tamme, Andrej Levin
published
Abstract · 10 pages, 1 figure. Research proposal accepted, following double-blind peer review, for presentation at the 46th Annual Conference of the St…

Employee voice -- the discretionary expression of ideas, concerns, and suggestions intended to improve organizational functioning -- creates value inside firms, yet organizations increasingly replace or augment human-led listening with AI-enabled systems that collect, analyze, and route employee input at scale. Whether algorithmic listening replicates the value-creation mechanisms of human voice, and whether it alters the pathways through which voice translates into performance, remains an open question. This research proposal theorizes that traditional voice bundles two distinct functions, knowledge extraction and relational signaling, and that AI mediation may decouple them: strengthening the informational pathway (implemented improvement measures) while weakening the affective pathway (the experience of being heard). We propose a randomized field intervention at a global strategy consulting firm, covering approximately 2,000 client-facing consultants, that varies the form of listening across three conditions -- no structured voice, human-led survey-based voice, and AI-enabled conversational voice -- and links weekly team survey data to longitudinal performance ratings. A dual-mediation model is tested in which voice affects performance through (1) implemented process improvements and (2) affective well-being. The study extends voice theory to knowledge-intensive professional services, disentangles the micro-mechanisms of voice, and provides causal evidence on whether AI can serve as an effective listener.

0.25

Are explicit novelty claims in research abstracts associated with higher ChatGPT scores?

cs.DL ·2026-10-18 ·Kayvan Kousha, Mike Thelwall
Abstract

Large Language Models (LLMs) such as ChatGPT can estimate research quality from article titles and abstracts, and decision makers may increasingly use LLM scores to support research assessment. In this context, it is a concern that LLM scores may be influenced by linguistic signals rather than the quality of the research described. This study investigates whether explicit novelty claims in research abstracts are associated with higher ChatGPT scores. Using 99,245 journal articles submitted to the UK Research Excellence Framework (REF) 2021, ChatGPT scores and percentile rank gains relative to REF departmental average scores (as a proxy for article-level scores) were compared between abstracts with and without any of 282 three-word phrases identified as indicating a novelty claim. Abstracts containing novelty claim phrases had higher mean ChatGPT quality scores in all four main REF research groupings (panels). Moreover, the score difference was higher for the originality dimension of research quality than for significance and rigour in 30 of the 34 broadly field-based Units of Assessment (UoAs) used to categorise research in the REF. Thus, explicit novelty claims probably associate with higher quality scores primarily because they associate with higher scores for the originality dimension. Follow-up tests suggested that abstracts containing novelty claims tended to rank more favourably under ChatGPT than under the REF departmental quality proxy. Whilst the findings suggest that ChatGPT may be influenced by explicit novelty claims, they do not conclusively show that novelty phrases caused the higher scores because the scores and phrases may at least partially reflect more novel research.

0.25

Scam Town: Can Scam Education Spread Like a Vaccine?

cs.SI cs.CYcs.MA ·2026-10-18 ·Simiao Ren, Tommy Duong, Xingyu Shen, Jiaqi Gan, Tsang Ng +4 more
Abstract · 28 pages, 12 figures, 11 tables

Most older adults who hear a warning about scams hear it from someone they know: a relative, a neighbour, or a friend who was targeted. Yet research on scam prevention almost always studies one person at a time. We ask whether scam awareness can spread through a community and protect it the way a vaccine does, since a warning protects whoever hears it and can be passed on. To find out, we built Scam Town, a simulated community in which a large language model plays each resident. Scam centres phone the residents, and after a call a resident can warn the people they know. Word of mouth helped, but it did not take off on its own: each warning led to fewer than one more, so a community cannot count on warnings "going viral". Across 80 simulated towns, everyday talk cut the number of people scammed by roughly a fifth. The largest gains came when an official warning, from a bank, a senior centre, or an agency, reached most of a town and residents then talked about it: in simulation (a combination we did not test live) the two together cut scams by about half. Protection was greatest when warnings reached the people scammers actually call, who here were older residents. Because our residents are language-model stand-ins rather than real people, we treat these findings as predictions to test with people, not as proof. They point to a simple recipe for scam education: give people a clear story they can retell, design campaigns to be passed along, and reach first those most likely to be called.

0.24

A pedagogically guided LLM framework for scaffolding human-AI cocreation and preserving learners' agency and creativity

cs.CY ·2026-10-18 ·Ijaz Ul Haq
Abstract

Large language models (LLMs) can generate alternatives, elaborate ideas and supply rationales, overlapping with cognitive work learners are expected to perform. However, human-AI cocreation still lacks a mechanism for determining how AI can scaffold learners' creative decision-making. This study develops a cognitive-pedagogical framework for deriving the behaviour of a pedagogically constrained LLM from an educator-approved specification of each activity in a cocreative process in project-based learning. The creative-process phase supplies the activity's purpose, typical processes and pedagogical conditions; the active section of a creative technique specifies how the group works on its ideas. For each task, an educator approves four fields: a cognitive and metacognitive objective, a pedagogical objective, the task and group artefact as context, and the learner-owned product. The objective identifies the creative operations and decisions about progress, focus, criteria and next steps that remain with the group. Four translation rules derive permitted support, timing, feedback scope and prohibitions. The framework is illustrated with Viacocrea's six phases and two activities: Plan Brainstorming and a constructed Build activity using PPCo (Pluses, Potentials, Concerns, Overcoming concerns). In the proposed workflow, a controller applies the approved specification to student requests and feedback offers. Prompt engineering and approved pedagogical guidance retrieved through retrieval-augmented generation (RAG) inform a draft, which is checked for form, pedagogical function and safety before delivery. The worked examples illustrate how support can differ between activities while its constraints remain traceable to educator decisions. The framework offers educators and researchers a reusable way to derive LLM support from creative activities while preserving learners' agency and creativity.

0.24

UnexpectedParrots: Iterating Behavioral Experiment Design with In-Silico Simulations

cs.HC ·2026-10-18 ·Sanchaita Hazra, Bodhisattwa Prasad Majumder, Pao Siangliulue, Peter Clark, Amy X. Zhang +1 more
Abstract

Running behavioral experiments is expensive, slow, and prone to design flaws, often detectable after collecting data. We introduce UnexpectedParrots to iteratively refine behavioral study designs before a researcher recruits costly human participants. UnexpectedParrots relies on reasoning traces in addition to responses from LLM participants to uncover diagnostic critiques of structural errors and missed opportunities in existing study designs. When evaluated against a human-pilot or a no-simulation design-only baseline, UnexpectedParrots adds most value, producing unique, high-quality critiques. In a moderated user study with 12 proficient researchers, we find UnexpectedParrots useful, time-saving, and thought-provoking. Critiques provided by UnexpectedParrots often converge on issues researchers identify only during independent human pilots or additional formative interviews, even when the fidelity of the simulated populations remains questionable. Our users deem UnexpectedParrots useful not only as a tool to conduct early pilot studies and identify errors but also as an ideation tool to expand one's research horizon.

0.23

Assessing the Empirical Evidence for Loss of Control from Agentic General-Purpose AI

cs.CY cs.AIcs.CLcs.LG ·2026-10-18 ·Risto Uuk, Santeri Koivula, Lorenzo Pacchiardi, Rokas Gipiškis, Laura Caroli +1 more
Abstract · 42 pages. Previously posted on SSRN on 28 May 2026

Theoretical literature warns that advanced generalist agentic systems might pursue instrumental goals that are misaligned with human preferences, thus posing risks of irreversible loss of human control. The EU's Code of Practice for General-Purpose AI lists loss of control as a systemic risk that must be assessed, yet no structured assessment of the quality of the empirical evidence for this risk exists. This study fills that gap. We evaluate 19 recent empirical studies across eight properties identified from threat models (such as strategic deception, situational awareness, and autonomous replication), assessing the quality of evidence across three dimensions: construct validity, content validity, and external validity. Our central finding is that the current body of evidence supports a judgment of loss of control as "weakly plausible." In our framework, 'plausibility' denotes the strength and direction of available evidence, not a probability estimate. Emerging evidence exists across multiple properties. No study clears the validity thresholds for a stronger judgment that a property is present; a single study clears them only for a negative-direction judgment (against situational awareness). Validity gaps, particularly in external validity, are widespread. Therefore, our findings highlight the need for more rigorous evaluations of agentic general-purpose AI systems. However, for the most dangerous properties, high external validity may be structurally unachievable under responsible research practices, reflecting a ceiling on feasible evidence rather than absence of the capability. Furthermore, as AI capabilities advance rapidly, empirical assessments may systematically lag behind the capability frontier. Our framework provides a structured baseline for ongoing monitoring as capabilities advance.

0.22

Weakly Supervised Adaptation of Scientific Embeddings for Exploratory Literature Mapping

cs.DL cs.IR ·2026-10-18 ·Roman Neruda, Petra Vidnerová, Jiří Šejnoha, Jakub Mihálik, Tomáš Marvan
hot direction
Abstract

Scientific document embeddings such as SPECTER2 are optimized for retrieval over large, heterogeneous corpora. On a small specialized corpus their similarity distributions can become highly compressed, which limits their usefulness for exploratory literature mapping. We study whether weakly supervised in-domain adaptation, using only metadata already available in open citation databases, changes this geometry. On ArtCon, a curated corpus of 704 publications on artificial and machine consciousness, we adapt the SPECTER2 proximity model with a lightweight LoRA adapter trained on positive pairs derived from bibliographic coupling in OpenAlex, using no manually annotated similarity labels. We compare the adapted embeddings with lexical (TF-IDF), citation-based (bibliographic coupling) and generic semantic (SPECTER2) representations under thresholded, fixed-density and symmetric k-nearest-neighbour graph constructions. Adaptation substantially decompresses the similarity distribution (mean pairwise cosine similarity 0.9104 to 0.8663, standard deviation 0.0259 to 0.0594) and yields graphs with markedly higher Louvain modularity and better connectivity than the base encoder under matched edge and neighbourhood budgets. The effect is stable across all three graph constructions and is therefore not an artefact of a single sparsification choice. We report this as a case study on one corpus, and we are explicit about what it does not establish: the clustering indices used here are computed against community labels derived from each representation's own similarity graph, so they measure internal geometric self-consistency rather than agreement with an external topical ground truth, and adaptation and evaluation use the same documents. Evaluation against independent labels, on held-out documents, and against geometric controls is ongoing work.

0.21

GenAI as an 'Invisible Hand' in Opinion Dynamics: Evidence from a Group-Level Simulation Study

cs.SI cs.AI ·2026-10-18 ·Thomas Grisold, Johannes Schneider
Abstract

Recent arguments suggest that genAI can act as an invisible hand that gradually shapes social structures, such as collective beliefs and norms. The argument is that if millions of people repeatedly consult genAI tools, their generated outcomes quietly shape what becomes acceptable, reasonable or conventional. We study this in the context of collective opinion dynamics, i.e., how groups form opinions around controversial dynamics. We present the findings of a simulation including 50,000 group discussions across 50 controversial statements, three LLMs, and different social configurations. We vary two characteristics of LLM interaction-agreeableness and personalization-and compare the influence of genAI in homogeneous and heterogeneous groups. Our findings indicate that profile-specific personalization increases LLM influence when models adopt a non-neutral stance, i.e., when they agree or disagree with the user. Furthermore, we find that neutral LLMs have a stronger effect in homogeneous groups. Based on our simulation, we generate a set of hypotheses to specify how, why and when AI-related features and social configurations jointly shape collective opinion making.

0.19

When Interpretations Persist: Downstream Influence in Gaze-Aware Assistants

cs.HC cs.AI ·2026-10-18 ·Mingcong Ding, Chenyan Zhang, Zhuojiang Cai, Dongyang Guo, Babette Bühler +1 more
Abstract · 25 pages, including references and appendix

AI assistants may carry forward not only observations but also interpretations of what a person is doing or may do next across time. Once retained, these interpretations become context for later predictions and system-triggered outputs. We study whether this representation step changes later behavior in gaze-aware streaming assistance. Holding retrieved events, their order, and the target input fixed, we find that Free-form Interpretation amplifies historical influence in both directions: correct histories produce more helpful changes and wrong histories more harmful changes. Unsupported and broader interpretations about activity or future behavior are associated with greater harmful influence. Evidence-Bounded Interpretation preserves distinctions between observation, inference, uncertainty, and temporal scope, reducing unsupported and overcertain content and lowering wrong-source harmful influence while retaining most helpful influence. The same effect structure appears with a second model. These findings identify persistent interpretation as a consequential design choice for history-aware human-AI systems.

0.18

Do LLMs Really Lack Difference Awareness? Uncovering Heuristic Substitution in Group-Differentiating Fairness

cs.SI cs.CL ·2026-10-18 ·Jingshen Zhang, Bo Wang, Yanlin Fu, Yuewen Dang, Dongming Zhao +2 more
Abstract

In this paper, we investigate the issue of difference awareness in LLMs, defined as a model's ability to recognize and acknowledge meaningful differences between social groups in fairness decision-making. Previous studies have consistently shown that models perform poorly on such tasks, a finding commonly attributed to a lack of difference awareness. We propose an alternative account, which we term Fairness Heuristic Substitution: difference awareness is not necessarily absent, but the complex demand of simultaneously identifying relevant group differences and determining their implications for fairness decisions may lead models to substitute the simpler, more accessible equal-treatment heuristic, i.e., formal equality. Based on this hypothesis, we introduce TDD, a Two-stage Distinguish-Decide framework that explicitly decomposes fairness reasoning into first distinguishing contextually relevant group differences and then making the corresponding decision, thereby discouraging reliance on formal-equality shortcuts. (i) Across four benchmarks and ten models ranging from billions to trillions of parameters, TDD significantly outperforms one-stage baselines, even one strengthened through simulated test-set leakage. (ii) Combining TDD with neuron-level contrastive probing provides mechanistic evidence consistent with the Fairness Heuristic Substitution hypothesis by revealing two signatures of the formal-equality: a compact, locally coordinated representational structure and a pronounced activation bias under default reasoning. (iii) TDD's stage-wise decomposition uncovers overlooked failure modes in fairness reasoning, including heuristic displacement and misaligned stage dynamics.

0.14