Out-Of-Distribution Generalization on Graphs: Paper List

Summarized by Lab of Media and Network, Department of Computer Science and Technology, Tsinghua University

Maintainer: Haoyang Li

Overview

Paper list of Graph Out-of-Distribution Generalization. The existing literature can be summarized into three categories from conceptually different perspectives, i.e., data, model, and learning strategy, based on their positions in the graph machine learning pipeline. For more details, please refer to our survey paper Out-Of-Distribution Generalization on Graphs: A Survey.

Data

Graph Data Augmentation

[CVPR 2023] Mind the Label Shift of Augmentation-based Graph OOD Generalization [paper]
[arXiv 2022] Adversarial Causal Augmentation for Graph Covariate Shift [paper]
[NeurIPS 2022] Knowledge Distillation Improves Graph Structure Augmentation for Graph Neural Networks [paper]
[ICML 2022] G-Mixup: Graph Data Augmentation for Graph Classification [paper]
[ICML 2022] Local Augmentation for Graph Neural Networks [paper]
[KDD 2022] Graph Rationalization with Environment-based Augmentations [paper]
[CVPR 2022] Robust Optimization as Data Augmentation for Large-scale Graphs [paper]
[NeurIPS 2021] Metropolis-Hastings Data Augmentation for Graph Neural Networks [paper]
[AAAI 2021] Data Augmentation for Graph Neural Networks [paper]
[NeurIPS 2020] Graph Random Neural Network for Semi-Supervised Learning on Graphs [paper]

Model

Disentanglement-based Graph Models

[TKDE 2022] Disentangled Graph Contrastive Learning With Independence Promotion [paper]
[NeurIPS 2022] Debiasing Graph Neural Networks via Learning Disentangled Causal Substructure [paper]
[NeurIPS 2021] Disentangled Contrastive Learning on Graphs [paper]
[AAAI 2020] Independence Promoted Graph Disentangled Networks [paper]
[NeurIPS 2020] Factorizable Graph Convolutional Networks [paper]
[KDD 2020] Interpretable deep graph generation with node-edge co-disentanglement [paper]
[ICML 2019] Disentangled Graph Convolutional Networks [paper]

Causality-based Graph Models

[TKDE 2022] OOD-GNN: Out-of-Distribution Generalized Graph Neural Network [paper]
[NeurIPS 2022] OOD Link Prediction Generalization Capabilities of Message-Passing GNNs in Larger Test Graphs [paper]
[NeurIPS 2022] Learning Causally Invariant Representations for Out-of-Distribution Generalization on Graphs [paper]
[ICML 2022] Learning from Counterfactual Links for Link Prediction [paper]
[KDD 2022] Causal Attention for Interpretable and Generalizable Graph Classification [paper]
[arXiv 2022] Deconfounding to Explanation Evaluation in Graph Neural Networks [paper]
[TNNLS 2022] Debiased Graph Neural Networks with Agnostic Label Selection Bias [paper]
[arXiv 2021] Generalizing Graph Neural Networks on Out-Of-Distribution Graphs [paper]
[ICML 2021] Generative Causal Explanations for Graph Neural Networks [paper]
[ICML 2021] Size-Invariant Graph Representations for Graph Classification Extrapolations [paper]

Learning Strategy

Graph Invariant Learning

[NeurIPS 2022] Learning Invariant Graph Representations for Out-of-Distribution Generalization [paper]
[NeurIPS 2022] Dynamic Graph Neural Networks Under Spatio-Temporal Distribution Shift [paper]
[NeurIPS 2022] SizeShiftReg: a Regularization Method for Improving Size-Generalization in Graph Neural Networks [paper]
[ICML 2022] Interpretable and Generalizable Graph Learning via Stochastic Attention Mechanism [paper]
[ICLR 2022] Handling Distribution Shifts on Graphs: An Invariance Perspective [paper]
[ICLR 2022] Discovering Invariant Rationales for Graph Neural Networks [paper]
[NeurIPS 2021] Shift-Robust GNNs: Overcoming the Limitations of Localized Graph Training Data [paper]
[arXiv 2021] Stable Prediction on Graphs with Agnostic Distribution Shift [paper]

Graph Adversarial Training

[AAAI 2023] Adversarial Weight Perturbation Improves Generalization in Graph Neural Network [paper]
[arXiv 2021] CAP: Co-Adversarial Perturbation on Weights and Features for Improving Generalization of Graph Neural Networks [paper]
[arXiv 2021] Distributionally Robust Semi-Supervised Learning Over Graphs [paper]
[arXiv 2021] Online Adversarial Distillation for Graph Neural Networks [paper]
[TKDE 2019] Graph Adversarial Training: Dynamically Regularizing Based on Graph Structure [paper]
[ICDM 2019] Domain-Adversarial Graph Neural Networks for Text Classification [paper]

Graph Self-supervised Learning

[arXiv 2022] Test-Time Training for Graph Neural Networks [paper]
[arXiv 2022] GraphTTA: Test Time Adaptation on Graph Neural Networks [paper]
[ICML 2022] Let Invariant Rationale Discovery Inspire Graph Contrastive Learning [paper]
[WWW 2022] Confidence May Cheat: Self-Training on Graph Neural Networks under Distribution Shift [paper]
[arXiv 2021] Graph Self-Supervised Learning: A Survey [paper]
[ICML 2021] From Local Structures to Size Generalization in Graph Neural Networks [paper]
[NeurIPS 2020] Graph Contrastive Learning with Augmentations [paper]
[ICLR 2020] Strategies for Pre-training Graph Neural Networks [paper]

Theory

[NeurIPS 2022] Generalization Analysis of Message Passing Neural Networks on Large Random Graphs [paper]
[NeurIPS 2021] Subgroup Generalization and Fairness of Graph Neural Networks [paper]
[NeurIPS 2021] Learning Theory Can (Sometimes) Explain Generalisation in Graph Neural Networks [paper]
[ICLR 2021] How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks [paper]
[ICLR 2021] A pac-bayesian approach to generalization bounds for graph neural networks [paper]
[arXiv 2021] Generalization bounds for graph convolutional neural networks via Rademacher complexity [paper]
[ICML 2021] Graph Convolution for Semi-Supervised Classification Improved Linear Separability and Out-of-Distribution Generalization [paper]
[ICML 2020 WorkShop] From Graph Low-Rank Global Attention to 2-FWL Approximation [paper]
[ICML 2020] Generalization and Representational Limits of Graph Neural Networks [paper]
[NeurIPS 2019] Graph Neural Tangent Kernel: Fusing Graph Neural Networks with Graph Kernels [paper]
[KDD 2019] Stability and Generalization of Graph Convolutional Neural Networks [paper]
[Neural Networks] The Vapnik–Chervonenkis dimension of graph and recursive neural networks [paper]

GNN Architecture

[arXiv 2022] Towards Better Out-of-Distribution Generalization of Neural Algorithmic Reasoning Tasks [paper]
[ICML 2022] Graph Neural Architecture Search Under Distribution Shifts [paper]
[arXiv 2021] Learning to Pool in Graph Neural Networks for Extrapolation [paper]
[ICLR 2020] What Can Neural Networks Reason About? [paper]
[ICLR 2020] Neural Execution of Graph Algorithms [paper]
[NeurIPS 2019] Understanding Attention and Generalization in Graph Neural Networks [paper]
[arXiv 2020] Customized Graph Neural Networks [paper]

Dynamic Environment

[NeurIPS 2022] Association Graph Learning for Multi-Task Classification with Category Shifts [paper]
[IJCNN 2021] Lifelong Learning of Graph Neural Networks for Open-World Node Classification [paper]

Domain Knowledge

[NeurIPS 2022] Learning Substructure Invariance for Out-of-Distribution Molecular Representations [paper]
[arXiv 2022] A critical examination of robustness and generalizability of machine learning [paper]
[TMLR 2022] How Do Graph Networks Generalize to Large and Diverse Molecular Systems [paper]
[AAAI 2022] How Does Knowledge Graph Embedding Extrapolate to Unseen Data: a Semantic Evidence View [paper]
[NeurIPS 2021 Workshop] Reliable Graph Neural Networks for Drug Discovery Under Distributional Shift [paper]
[ICML 2020 Workshop] Evaluating Logical Generalization in Graph Neural Networks [paper]

Adaptation

[arXiv 2022] Empowering Graph Representation Learning with Test-Time Graph Transformation [paper]
[arXiv 2022] Shift-Robust Node Classification via Graph Adversarial Clustering [paper]
[KDD 2022] Learning on Graphs with Out-of-Distribution Nodes [paper]

Dataset

[NeurIPS 2022] GOOD: A Graph Out-of-Distribution Benchmark [paper]
[arXiv 2022] DrugOOD: Out-of-Distribution (OOD) Dataset Curator and Benchmark for AI-aided Drug Discovery -- A Focus on Affinity Prediction Problems with Noise Annotations [paper]
[NeurIPS 2021 Workshop] A Closer Look at Distribution Shifts and Out-of-Distribution Generalization on Graphs [paper]

Others

[arxiv 2022] Towards Better Generalization with Flexible Representation of Multi-Module Graph Neural Networks [paper]
[NeurIPS 2022] GraphDE: A Generative Framework for Debiased Learning and Out-of-Distribution Detection on Graphs [paper]
[arXiv 2022] Causally-guided Regularization of Graph Attention Improves Generalizability [paper]