CS Paper Radar

Graph Label Alignment: A Diagnostic Atlas for Graph Classification

Neelam Akula, Varun Shiralkar, Murat Kantarcioglu, Baris Coskunuzer

cs.LG cs.AI ·submitted 2026-11-10 21:36 UTC · v1
top venuebenchmark

Abstract

Graph classification benchmarks vary widely in difficulty: some are solved by simple pooled baselines, while others remain challenging for GNNs and graph transformers. We take a data-centric view and ask whether benchmark success is driven by genuine structure--feature reasoning or by label signal already accessible from restricted marginal views. We introduce graph label alignment, a lightweight diagnostic that measures label accessibility through two supervised probes computed without training a graph encoder: structural label alignment (SLA), based on permutation-invariant topological and spectral summaries, and feature label alignment (FLA), based on pooled node features with edges removed. Across real graph classification benchmarks and multiple pooled backbones, the resulting SLA--FLA plane strongly organizes test accuracy, with low-alignment datasets forming a consistent regime where standard pooled pipelines struggle. To probe what lies beyond marginal accessibility, we introduce controlled interaction-dominated stress tests, including heterophilic ego-graphs and synthetic Twins tasks whose structural and feature marginals are matched across classes but whose role--feature couplings differ. These results show that high alignment explains much of easy benchmark performance, while low alignment reveals regimes where success requires preserving structure--feature correspondence beyond standard pooled representations.

Comment
40 pages
Journal ref
NeurIPS 2026

Closest papers

Nearest neighbours by meaning across the whole index.

Implicit Feature Reweighting for Label-Free Graph Generalization

cs.LG ·2026-10-21 ·Tianjun Yao, Chen Xi, Xinyang Li, Lixian Zhang
code/data similarity 0.68
Abstract · We propose an implicit feature reweighting method to enhance graph-level OOD generalization without class labels

Graph-level out-of-distribution (OOD) generalization is critical for deploying graph neural networks in real-world applications where distribution shifts are inevitable. Existing approaches typically require class labels during representation learning to decide which substructures are predictive, limiting their applicability. We propose SCARCE, a label-free representation-learning framework that yields distribution-shift-robust graph representations from an environment partition of the training graphs alone, whether that partition comes from dataset metadata or, when none is available, from clustering a self-supervised embedding. Our method operates in two stages: (1) a variational graph autoencoder is warm-started for graph reconstruction; and (2) an adversarial confusion mechanism forces the encoder to produce representations from which a discriminator cannot distinguish environments, while preserving graph-reconstruction fidelity. We provide theoretical analysis showing that environment confusion contracts environment-discriminative directions in the representation space, which, under regularized downstream classifiers, is equivalent to imposing a stronger directional penalty on environment-associated features. This feature-reweighting mechanism reduces reliance on spurious correlations without erasing information, offering a label-free, continuous relaxation of last-layer reweighting methods. Experiments on a controlled synthetic datasets and the DrugOOD datasets show that SCARCE achieves state-of-the-art OOD generalization among label-free methods, using no class labels during representation learning.

0.22

Complementary Supervised and Self-Supervised Representations for Out-of-Distribution Graph Learning

cs.LG ·2026-10-06 ·Qingying Hao, Zikang Chen, Chuxuan Hu, Jinyuan Jia, Bo Li +2 more
similarity 0.68
Abstract

Out-of-distribution (OOD) generalization remains challenging for graph neural networks (GNNs), as graph distributions can vary substantially across time and domains. Supervised and self-supervised graph representation learning are guided by distinct objectives and offer different perspectives on graph representations. In this work, we study whether self-supervised representations (SSL) can provide complementary signals to improve supervised OOD node classification. We develop two backbone-agnostic frameworks that exploit such information at different stages of learning and prediction. Co-Train jointly learns supervised and SSL representations and adaptively integrates them during training, while Dual-Space Retrieval performs non-parametric prediction in the two representation spaces and combines their predictions through confidence-aware fusion at inference time. The supervised and SSL encoders are separately parameterized and need not share the same GNN architecture. We evaluate multiple GNN backbones and two distinct SSL objectives, DGI and GRACE, on four graph benchmarks spanning temporal and cross-domain distribution shifts. Extensive experiments show that Co-Train consistently outperforms strong supervised OOD baselines, while Dual-Space Retrieval achieves competitive performance as a flexible non-parametric alternative. Results across different backbones and SSL objectives, together with representation analyses and ablations, demonstrate that SSL representations provide complementary information to supervised representations and can improve OOD node classification across diverse settings.

0.19

Alleviating Limited Effectiveness and Multi-Class Collapse in Deep Graph-level Clustering

cs.LG ·2026-11-04 ·Xiaoyang Lin, Renchi Yang, Yuqi Chen, Xin Huang
benchmarktheoryhot direction similarity 0.67
Abstract

Given a collection of $N$ graphs, graph-level clustering (GLC) seeks to partition the $N$ graphs into semantically meaningful groups, which serves as a fundamental task in analyzing graph-structured data. While Graph Isomorphism Networks (GINs) have become the de facto encoders for GLC, they still fail to distinguish graphs due to their neglect of global topology. On top of that, existing deep clustering frameworks suffer from severe "clustering collapse" issues, particularly in imbalanced multi-class scenarios. To address these challenges, we propose S$^3$GLC, a novel framework that integrates global topological characteristics with local neighborhood features for balanced GLC. Theoretically, we leverage the Hoffman-Wielandt inequality to construct Spectral Signature as a rigorous lower bound for graph edit distance (GED), providing a computationally efficient proxy for global topological features. We additionally propose a joint manifold-preserving sparse regression module to fuse Spectral Signature with Spatial Representation, eliminating noise while preserving the complementary information. Furthermore, to mitigate clustering collapse, we introduce a balanced clustering objective based on optimal transport, harnessing the Sinkhorn-Knopp normalization to enforce equipartition constraints on cluster assignments. Extensive experiments on 10 real-world datasets demonstrate that S$^3$GLC significantly outperforms 15 state-of-the-art baselines, achieving remarkable improvements on benchmark datasets such as REDDIT-12K and OGBG-PPA.

0.23

The Universal Classifier for Graph Learning

cs.LG ·2026-09-28 ·Ben Finkelshtein, André Linhares, Petar Veličković, Bryan Perozzi, Mikhail Galkin
similarity 0.66
Abstract

While foundation models have revolutionized natural language processing and computer vision by leveraging universal vocabularies, Graph Machine Learning (GML) remains fractured due to the absence of a unified feature and structural representation across diverse domains. Existing works claiming to be Graph Foundation Models (GFMs) are typically restricted to node-level predictions or require fixed feature dimensions, failing to provide a truly task-agnostic backbone for the full spectrum of graph learning applications. In this paper, we introduce the Universal Classifier (UC), which supports arbitrary feature and class cardinalities, unifying node-, edge-, and graph-level objectives under a single similarity-based classification objective. The UC reformulates all node-, edge-, and graph-level prediction tasks as maximizing similarity in the latent space: by lifting heterogeneous features and labels into 3D latent tensors, the model learns transferable features independent of specific input schemas. This architecture allows a single pre-trained model to generalize to node classification, node regression, and link prediction across unseen graphs with varying feature semantics. Experiments show strong zero-shot transfer performance across node-, link-, and graph-level tasks.

0.32

Graph-dependent pooling: adaptive granularity in hierarchical graph neural networks

cs.AI cs.LG ·2026-10-24 ·Bohan Yoon, Hyejin Jang
similarity 0.66
Abstract

Graph Neural Networks are widely used to learn structural representations of relational data. Hierarchical graph pooling learns multi-level representations by progressively coarsening graphs; however, prior methods rely on predefined cluster numbers or fixed pooling ratios, limiting adaptability to individual graphs. We propose Graph-Dependent Pooling (GDPool), which dynamically determines pooling granularity for each graph. We analyze optimization conflicts among layer-wise objectives through gradient alignment, showing that independently supervised pooling levels can send conflicting gradients to shared parameters. GDPool uses self-attention to select graph-specific anchors representing structurally important nodes and constructs coarse graphs through soft node assignment. Experiments on metal-organic framework and molecular graph benchmarks show that GDPool consistently achieves the best predictive performance across all benchmarks. Evaluations using modularity and the effective number of clusters show that GDPool achieves consistently higher modularity than existing methods while the number of active clusters decreases monotonically with depth. Sensitivity analyses confirm stable behavior across hyperparameter and architectural choices.

0.13

Message Passing Does More with Less for In-Context Learning on Graphs

cs.LG cs.SI ·2026-09-29 ·Dooho Lee, Jinmo Lee, Minho Jeong, Kijung Shin, Jaemin Yoo
code/data similarity 0.66
Abstract

Achieving strong performance with graph neural networks (GNNs) typically requires training and hyperparameter tuning for each dataset, incurring repeated costs and effort. Graph in-context learning (ICL) avoids this by using a single pretrained model to predict unknown node labels directly from labeled context nodes. Existing approaches, however, rely on dense attention across nodes, making inference increasingly expensive as graphs grow. In this work, we present Ephris, a new graph in-context learner built on sparse message passing, scaling linearly with the number of node-feature entries and graph edges. Ephris is pretrained entirely on synthetic graphs generated from structural causal models with diverse graph structures and relational dynamics, exposing the model to varied dependencies among topology, features, and labels. We evaluate Ephris on 51 node-classification datasets against 15 extensively tuned GNNs and existing graph ICL methods under both high- and low-label train/validation/test splits. Across both settings, Ephris ranks first on all four aggregate measures: Elo, improvability, average rank, and accuracy. Its inference cost remains comparable to training a single GNN once, while being over 10 times faster than previous graph ICL models. Together, these results advance the performance-runtime Pareto frontier, demonstrating that strong graph ICL does not require dense attention. Code and model weights are available at https://github.com/nums-ai/ephris.

0.25

Hierarchical Graph Representation Learning with Pooling-Induced Substructures

cs.LG ·2026-11-10 ·Luca Sbicego, Xiaowen Dong, Dorina Thanou
top venuecode/data similarity 0.65
Abstract · Accepted to NeurIPS 2026

Hierarchical graph-level tasks that require reasoning over substructures that extend beyond local neighborhoods are posing challenges for standard Graph Neural Networks (GNNs). While graph pooling aims to address this by coarsening graphs into hierarchical representations, existing methods often fail to explicitly preserve the substructures they induce, leading to inconsistent empirical performance. Moreover, evaluating such approaches remains challenging due to the lack of benchmarks with explicit hierarchical structure. We propose Pooled-Substructure-Aware GNNs, an approach to graph pooling that explicitly preserves and processes substructures uncovered by structure-based pooling schemes. We show that this design enhances the expressive power of pooling-based GNNs in the sense of the Weisfeiler-Lehman (WL) test. To enable principled evaluation, we introduce two new synthetic datasets with controllable hierarchical structure and propose a metric to characterize the prevalence of hierarchical patterns in real-world graph-level benchmarks. On tasks with relevant hierarchical structure, our approach outperforms matched pooling baselines, while remaining competitive on structurally simpler benchmarks. Code and datasets are publicly available at https://github.com/lsbicego/PSA-GNN/.

0.43

Learning the Graph and the Embedding Together: Classifier-Independent Rewiring for Heterophilic Node Classification

cs.LG ·2026-09-26 ·Harshit Kumar, Sujan Chakraborty, Priyanka Saha, Pritam Kar, Saptarshi Bej
similarity 0.65
Abstract

Graph neural networks lose much of their advantage on heterophilic graphs, where connected nodes often carry different labels. Graph rewiring is a popular remedy, but rewiring methods are usually evaluated with a single classifier, which makes it hard to tell whether the gains come from the new topology or from that particular pairing. We propose an affinity-guided rewiring method that estimates the graph and the node representation together. It alternates, in the spirit of expectation maximisation, between training a lightweight graph neural network on the current graph and re-weighting candidate edges under a modularity objective with a pseudo-label homophily term. Candidate edges come from a compact pool scored by a contrastively learned node similarity and a neighbourhood-distribution affinity. The method returns two classifier-independent outputs: a rewired graph and a node embedding learned on it. Across six heterophilic benchmarks and five downstream classifiers, it improves accuracy over the original graph with normalised features in 23 of 30 classifier-dataset combinations, with a mean gain of 5.8 points, and reduces the accuracy spread between classifiers about fourfold. A controlled ablation shows that the two outputs are each useful and play complementary roles: the embedding contributes most of the accuracy gain, while the rewired graph makes different classifiers agree. A fully unsupervised variant, which uses no labels during rewiring, retains most of the improvement. The rewired graphs are also more homophilic and improve label propagation and community detection.

0.25