Temporal Memory-Aware Online Test-Time Adaptation on Dynamic Graphs
Machine Learning
Summary
The authors address the problem of making graph neural networks (GNNs) work better when testing on new, changing graphs, especially those where connections and node information change over time. They propose a method called DGOTTA that adapts these models during testing by generating diverse graph versions, using memory to avoid forgetting what was learned, and keeping predictions consistent over time. Their approach helps models handle different types of changes in dynamic graphs and improves test-time performance across various datasets and model types.
Authors
Bo Li, Xin Zheng, Ming Jin, Can Wang, Shirui Pan
Abstract
Test-time adaptation (TTA) on graphs aims to adapt a graph neural network (GNN) that is well-trained on the training graph to the test graph, which involves potential distribution shifts that may harm model generalization and test-time inference. While recent efforts have investigated TTA on static graphs, there is still a research gap on dynamic graphs learned with dynamic GNN (DGNN) models, where both structural connectivity and node semantics evolve continuously over time. This makes adapting a DGNN model for reliable test-time performance substantially challenging. To fill this gap, in this work, we propose a novel framework of temporal memory-aware Online Test-Time Adaptation on Dynamic Graphs, named DGOTTA, to effectively adapt well-trained DGNNs during test time. Specifically, the proposed DGOTTA contains three modules: (1) temporal-aware augmentation, to extend the diversity of test dynamic graphs for addressing complex temporal and spatial shifts; (2) memory-aware model prediction, to alleviate catastrophic forgetting; (3) consistency-guided online adaptation, to enforce temporal alignment and memory smoothness. Extensive experiments on three real-world datasets and four DGNN backbones demonstrate that DGOTTA significantly improves generalization under diverse distribution shifts and multiple model architectures.