arXiv：cs.AI· Mahtab Sarvmaili·· 4 小时前
MPGE：面向分子分类解释的多视角图解释器
MPGE: A Multi-Perspective Graph Explainer for Molecular Classification Explanation
AI 导读
MPGE 是一种面向冻结 GNN 分类器的多视角图解释器，统一了事实支持（PT）、反事实敏感（CF）与样例容忍（EXE）三种解释视角。在 MUTAG、Mutagenicity、AIDS、COX2_MD 和 BBBP 上，成功事实掩码平均保留输入边的 8.6%–15.5%，有界反事实覆盖率为 4.8%–67.6%，样例保留覆盖率达 98.9%–100.0%。
正文
Abstract:Graph neural networks (GNNs) predict molecular properties from chemical graph data, but predictive accuracy does not explain how graph information supports an individual decision. A compact prediction-preserving rationale does not necessarily reveal which changes reverse the decision or which modifications the model tolerates. We propose the Multi-Perspective Graph Explainer (MPGE), unifying factual support, counterfactual sensitivity, and exemplar tolerance for a frozen classifier. The factual view, originally termed prototype (PT), seeks a compact retained edge set with the same label and required confidence. Counterfactual (CF) explanations seek bounded prediction-changing deletions; exemplar (EXE) explanations seek non-trivial bounded deletions that preserve the label and confidence. A shared constrained formulation connects prediction behavior, compactness, and edit cost, while separate objectives generate the three views. Our graph-classification extension of CF-GNNExplainer learns symmetric edge rankings and verifies discrete candidates, recording unsuccessful searches. A separate BBBP fragment backend returns RDKit-sanitized molecules. We evaluate the primary GCN implementation on MUTAG, Mutagenicity, AIDS, COX2_MD, and BBBP using semantic coverage, conditional quality, stability, and runtime. Successful factual masks retained 8.6%--15.5% of input edges on average across datasets; bounded counterfactual coverage was 4.8%--67.6%, and exemplar preservation coverage was 98.9%--100.0%. Exploratory controls reveal the influence of hard projection and retained node information. Quantitative comparisons and molecular visualizations characterize model support, sensitivity, and tolerance without treating them as validated chemical mechanisms.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.12039 [cs.LG] |
| (or arXiv:2610.12039v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.12039 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Mahtab Sarvmaili [view email]
[v1]
Thu, 8 Oct 2026 14:28:47 UTC (393 KB)
来源：arXiv：cs.AI · arxiv.org