Research map / Graph learning / GNN architectures
GStarX: Explaining Graph Neural Networks with Structure-Aware Cooperative Games
Shichang Zhang, Yozen Liu, Neil Shah, Yizhou Sun
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 structure-aware HN-value scoring instead of SubgraphX's Shapley-based explanation scoring | GCN explanation · BA2Motifs | H-Fidelity: 0.6050 -> 0.5824 | Hurt |
| Use structure-aware HN-value scoring instead of Shapley-based explanation scoring | GCN explanation · BACE | H-Fidelity: 0.5519 -> 0.5934 | Improved |
| Use structure-aware HN-value scoring instead of Shapley-based explanation scoring | GCN explanation · BBBP | H-Fidelity: 0.5610 -> 0.5227 | Hurt |
| Use structure-aware HN-value scoring instead of Shapley-based explanation scoring | GCN explanation · GraphSST2 | H-Fidelity: 0.5487 -> 0.5519 | Improved |
| Use structure-aware HN-value scoring instead of Shapley-based explanation scoring | GCN explanation · MUTAG | H-Fidelity: 0.5253 -> 0.6171 | Improved |
| Use structure-aware HN-value scoring instead of Shapley-based explanation scoring | GCN explanation · Twitter | H-Fidelity: 0.5494 -> 0.5716 | Improved |
| Use HN-value scoring rather than the strongest baseline | GAT explanation · GraphSST2 | H-Fidelity: 0.5484 -> 0.5542 | Improved |
| Use HN-value scoring rather than the strongest baseline | GIN explanation · MUTAG | H-Fidelity: 0.5592 -> 0.6064 | Improved |
| Use GStarX rather than SubgraphX | GCN explanation · BBBP | Time(s): 75.96 -> 31.24 | Improved |
Later papers that compared with this method
1 later accepted paper reports GStarX as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- SAME: Uncovering GNN Black Box with Structure-aware Shapley-based Multipiece Explanations (NeurIPS 2023 · better on 12 of 12 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- GOAt: Explaining Graph Neural Networks via Graph Output Attribution (ICLR 2024 · similarity 0.89)
- Self-Interpretable Graph Learning with Sufficient and Necessary Explanations (AAAI 2024 · similarity 0.89)
- Towards Robust Fidelity for Evaluating Explainability of Graph Neural Networks (ICLR 2024 · similarity 0.89)
- Explaining Graph Neural Networks via Structure-aware Interaction Index (ICML 2024 · similarity 0.89)
- Exact Computation of Any-Order Shapley Interactions for Graph Neural Networks (ICLR 2025 · similarity 0.89)
- Stratified GNN Explanations through Sufficient Expansion (AAAI 2024 · similarity 0.88)
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.
