Research map / Graph learning / Knowledge and temporal graphs
Explaining Temporal Graph Neural Network via Quantum-Inspired Evolutionary Algorithm
Masahiro Mitani, Yuya Sasaki
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use QIEA-TGX instead of GA-TGX / T-GNNExplainer / PGExplainer | TGAT and TGN explainers · Wikipedia, Reddit (factual and counterfactual) | fidelity: e.g. Reddit TGN counterfactual 0.122 (GA-TGX) -> 0.597 | Improved |
| use QIEA-TGX instead of GreeDy | TGAT/TGN counterfactual explanations · Wikipedia TGAT counterfactual | fidelity: 0.568 -> 0.560 (GreeDy higher) | No clear effect |
| use QIEA-TGX instead of GreeDy on runtime | explainers · Wikipedia, Reddit | runtime: up to 92% less computation time | Improved |
| add graph structure-aware optimization to QIEA-TGX (ablation) | QIEA-TGX · Reddit TGAT/TGN counterfactual | fidelity improvement: 24.4% (TGAT), 25.2% (TGN) | Improved |
| add graph structure-aware optimization to QIEA-TGX (ablation) | QIEA-TGX · Wikipedia factual | fidelity improvement: 0% / 0% | No clear effect |
| add graph structure-aware optimization to GA-TGX (ablation) | GA-TGX · Wikipedia and Reddit | fidelity improvement: -1.7% to 5.8% | Mixed |
| increase sparsity from 0.1 to 0.5 (ablation) | all explainers · Reddit TGAT | fidelity | Improved |
| use a very small H-gate threshold theta_H (ablation) | QIEA-TGX · Reddit | fidelity | Hurt |
Closest papers by meaning
- Training-free Counterfactual Explanation for Temporal Graph Model Inference (ICLR 2026 · similarity 0.91)
- CoDy: Counterfactual Explainers for Dynamic Graphs (ICML 2025 · similarity 0.87)
- OrphicX: A Causality-Inspired Latent Variable Model for Interpreting Graph Neural Networks (CVPR 2022 · similarity 0.87)
- Beyond Topological Self-Explainable GNNs: A Formal Explainability Perspective (ICML 2025 · similarity 0.85)
- COMRECGC: Global Graph Counterfactual Explainer through Common Recourse (ICML 2025 · similarity 0.85)
- EiG-Search: Generating Edge-Induced Subgraphs for GNN Explanation in Linear Time (ICML 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
