Pith. sign in

REVIEW 5 cited by

Discovering Invariant Rationales for Graph Neural Networks

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2201.12872 v1 pith:YG5DI3DT submitted 2022-01-30 cs.LG

classification cs.LG
keywords graphdatainterpretabilityinvariantrationalesbiasescausaldatasets
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Intrinsic interpretability of graph neural networks (GNNs) is to find a small subset of the input graph's features -- rationale -- which guides the model prediction. Unfortunately, the leading rationalization models often rely on data biases, especially shortcut features, to compose rationales and make predictions without probing the critical and causal patterns. Moreover, such data biases easily change outside the training distribution. As a result, these models suffer from a huge drop in interpretability and predictive performance on out-of-distribution data. In this work, we propose a new strategy of discovering invariant rationale (DIR) to construct intrinsically interpretable GNNs. It conducts interventions on the training distribution to create multiple interventional distributions. Then it approaches the causal rationales that are invariant across different distributions while filtering out the spurious patterns that are unstable. Experiments on both synthetic and real-world datasets validate the superiority of our DIR in terms of interpretability and generalization ability on graph classification over the leading baselines. Code and datasets are available at https://github.com/Wuyxin/DIR-GNN.

Discussion (0). Sign in to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. When, How Long and How Much? Interpretable Neural Networks for Time Series Regression by Learning to Mask and Aggregate

    cs.LG 2025-12 conditional novelty 7.0 of 10

    MAGNETS learns unsupervised, mask-based concepts to make time-series regression predictions additively interpretable, recovering ground-truth temporal rules on synthetic tasks and beating interpretable baselines on mo...

  2. OpFlow: Learning Opportunity-Conditioned Choice Potentials for Robust OD Flow Prediction

    cs.LG 2026-07 conditional novelty 6.5 of 10

    Robust OD flow prediction comes from learning row-centered exposure-to-choice potentials and reconstructing counts as scale times allocation, not from raw-count supervision.

  3. Cross-Resolution Semantic Learning for Graph Domain Adaptation

    cs.LG 2026-07 conditional novelty 6.0 of 10

    CReSL improves graph domain adaptation by learning cross-resolution source-to-target routing and grafting target representations toward source class prototypes.

  4. A Recipe for Causal Graph Regression: Confounding Effects Revisited

    cs.LG 2025-07 conditional novelty 6.0 of 10

    The paper proposes a contrastive-learning-based causal graph regression framework that explicitly models the predictive power of confounding subgraphs and achieves state-of-the-art OOD generalization on graph regressi...

  5. Do Explanations Increase the Risk of Decision Logic Leakage? Explanation-Guided Stealing of Graph Models

    cs.LG 2025-06 conditional novelty 6.0 of 10

    An attacker can exploit explanation heatmaps from deployed graph neural networks to train a surrogate that matches both predictions and highlighted decision logic, outperforming prior stealing attacks.

Pith tools