arXiv：cs.LG（机器学习，全量分类）· Dominik Matuszek, Bartosz Zieli\'nski, Tomasz Danel, Dawid Rymarczyk·· 14 小时前AI 评分34
WOMBAT：用于分子基准测试与归因测试的白盒预言机
WOMBAT: Whitebox Oracle for Molecular Benchmarking and Attribution Testing
AI 导读
研究者发布 WOMBAT，一个包含 14 个白盒 GNN 的基准，每个模型的消息传递权重由人工设定以检测特定 SMARTS 模体，从而提供归因真值。模型在数百万 PubChem 分子上验证，并评测了 GNNExplainer、PGExplainer 和 Integrated Gradients 等事后解释器。
正文
Abstract:When a graph neural network (GNN) explainer produces an unexpected attribution on a molecule, the attribution alone cannot reveal whether the explainer has failed or the model has learned a shortcut. We introduce WOMBAT, a benchmark of 14 whitebox GNNs, each with message-passing weights set by hand to detect a specific SMARTS motif. Each model's decision rule is known by construction, providing attribution ground truth against which explainer errors can be identified and studied. We validate the models on millions of PubChem molecules and evaluate post-hoc explainers including GNNExplainer, PGExplainer, and Integrated Gradients. Guided by our qualitative analysis, we construct a model that causes Integrated Gradients to spread attribution across the graph, even though the model reliably detects the intended motif. We release the dataset, models, and evaluation code to help researchers in the development of newer XAI tools for GNNs.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.00713 [cs.LG] |
| (or arXiv:2610.00713v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.00713 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Dominik Matuszek [view email]
[v1]
Wed, 30 Sep 2026 21:02:26 UTC (7,164 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org