Abstract:Dynamic graph neural networks (DyGNNs) have demonstrated powerful predictive abilities by exploiting graph structural and temporal dynamics. However, the existing DyGNNs fail to handle distribution shifts, which naturally exist in dynamic graphs, mainly because the patterns exploited by DyGNNs may be variant with respect to labels under distribution shifts. In this paper, we propose to handle spatio-temporal distribution shifts in dynamic graphs by discovering and utilizing {\it invariant patterns}, i.e., structures and features whose predictive abilities are stable across distribution shifts, which faces two key challenges: 1) How to discover the complex variant and invariant spatio-temporal patterns in dynamic graphs, which involve both time-varying graph structures and node features. 2) How to handle spatio-temporal distribution shifts with the discovered variant and invariant patterns. To tackle these challenges, we propose the Disentangled Intervention-based Dynamic graph Attention networks (DIDA). Our proposed method can effectively handle spatio-temporal distribution shifts in dynamic graphs by discovering and fully utilizing invariant spatio-temporal patterns. Specifically, we first propose a disentangled spatio-temporal attention network to capture the variant and invariant patterns. Then, we design a spatio-temporal intervention mechanism to create multiple interventional distributions by sampling and reassembling variant patterns across neighborhoods and time stamps to eliminate the spurious impacts of variant patterns. Lastly, we propose an invariance regularization term to minimize the variance of predictions in intervened distributions so that our model can make predictions based on invariant patterns with stable predictive abilities and therefore handle distribution shifts. Experiments on three real-world datasets and one synthetic dataset demonstrate the superiority of our method over state-of-the-art baselines under distribution shifts. Our work is the first study of spatio-temporal distribution shifts in dynamic graphs, to the best of our knowledge.

Invariant Node Representation Learning under Distribution Shifts with Multiple Latent Environments

Handling Distribution Shifts on Graphs: An Invariance Perspective

Towards Distribution Shift of Node-Level Prediction on Graphs: An Invariance Perspective

Graph Invariant Learning with Subgraph Co-mixup for Out-Of-Distribution Generalization

Spectral Invariant Learning for Dynamic Graphs under Distribution Shifts

IENE: Identifying and Extrapolating the Node Environment for Out-of-Distribution Generalization on Graphs

Leveraging Invariant Principle for Heterophilic Graph Structure Distribution Shifts

Graphs Generalization under Distribution Shifts

Dissecting the Failure of Invariant Learning on Graphs

Does Invariant Graph Learning via Environment Augmentation Learn Invariance?

Out-of-Distribution Generalized Dynamic Graph Neural Network with Disentangled Intervention and Invariance Promotion

Environment-Aware Dynamic Graph Learning for Out-of-Distribution Generalization

Discovering Invariant Neighborhood Patterns for Heterophilic Graphs

Fortune Favors the Invariant: Enhancing GNNs’ Generalizability with Invariant Graph Learning

Dynamic Graph Neural Networks Under Spatio-Temporal Distribution Shift

Invariant Graph Learning Meets Information Bottleneck for Out-of-Distribution Generalization

Learning Invariant Representations with a Nonparametric Nadaraya-Watson Head

Understanding the Generalizability of Link Predictors Under Distribution Shifts on Graphs

Learning Causally Invariant Representations for Out-of-Distribution Generalization on Graphs

When Invariant Representation Learning Meets Label Shift: Insufficiency and Theoretical Insights

Improving out-of-distribution generalization in graphs via hierarchical semantic environments