ST-TGExplainer: Disentangling Stability and Transition Patterns for Temporal GNN Interpretability
For researchers and practitioners using TGNNs, this work addresses the overlooked need to explain predictions based on both recurring and novel interactions, but the improvement is incremental over existing interpretable TGNNs.
The paper tackles the interpretability gap in temporal graph neural networks (TGNNs) by proposing ST-TGExplainer, a self-explainable model that disentangles stability (historical) and transition (first-time) patterns. Experiments show it achieves strong predictive performance and more faithful explanations, though no concrete numbers are provided.
Temporal graph neural networks (TGNNs) have gained significant traction for solving real-world temporal graph tasks. However, their interpretability remains limited, as most TGNNs fail to identify which historical interactions most influence a given prediction. Despite promising progress on interpretable TGNNs, existing methods predominantly focus on previously seen historical interactions, which we term stability patterns, while overlooking newly emerging first-time interactions, which we term transition patterns. Both types of patterns are essential for faithful temporal explanations. To address this limitation, we propose ST-TGExplainer, a self-explainable TGNN that disentangles Stability and Transition patterns in temporal graphs for a more faithful Temporal GNN Explainer. Guided by a disentangled information bottleneck objective, ST-TGExplainer learns a compact explanatory subgraph that remains predictive of the event label while explicitly suppressing label-conditioned redundancy between stability and transition patterns. Extensive experiments demonstrate that ST-TGExplainer achieves strong predictive performance and yields more faithful explanations. Code is available at https://github.com/hjchen-hdu/ST-TGExplainer.