ST-TGExplainer: Disentangling Stability and Transition Patterns for Temporal GNN Interpretability

2026-05-19

AI Summary

Existing approaches to interpreting temporal graph neural networks focus exclusively on historical recurrent interactions—referred to as stability patterns—while overlooking newly emerging first-time interactions, or transition patterns, resulting in incomplete explanations. This work proposes a self-explaining temporal graph neural network that explicitly disentangles and jointly models both stability and transition patterns for the first time. By introducing a disentangled information bottleneck objective, the method suppresses label-conditional redundancy between the two pattern types, thereby learning compact yet discriminative temporal explanatory subgraphs. Evaluated across multiple datasets, the proposed approach not only maintains state-of-the-art predictive performance but also significantly enhances explanation fidelity and interpretability.

Abstract

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.

Problem

• Temporal GNN interpretability
• Stability patterns
• Transition patterns
• Explainable AI
• Temporal graphs

Innovation

• Temporal GNN Interpretability
• Disentangled Explanation
• Stability Patterns
• Transition Patterns
• Information Bottleneck