发现GNN解释自不一致的原因并提出无需训练的修复方法
Why Self-Inconsistency Arises in GNN Explanations and How to Exploit It
- 识别出重解释导致的上下文扰动是自不一致主因
- 提出隐信号分配假说，解释为何仅部分边敏感于扰动
- 设计无需训练的去噪策略，提升解释质量且开销仅4-6%
近期研究发现，自解释图神经网络（SI-GNN）生成的解释可能存在自不一致：当模型再次作用于自身生成的子图时，会得到不同解释。本文首先指出重解释引发的上下文扰动是分数变化的直接原因。进而提出隐信号分配假说，解释为何仅部分边对扰动敏感，并分析简洁性正则化对隐信号分配的影响。由于自不一致边无法提供稳定证据，我们提出无需训练、无需微调的后处理方法Self-Denoising（SD），仅需一次额外前向传播即可校准解释。在多个代表性SI-GNN框架、主干网络和基准数据集上的实验验证了假设，结果表明SD能持续提升解释质量，实际计算开销仅增加4–6%。
原文摘要 · Abstract (English)
Recent work has observed that explanations produced by Self-Interpretable Graph Neural Networks (SI-GNNs) can be self-inconsistent: when the model is reapplied to its own explanatory graph subset, it may produce a different explanation. However, why self-inconsistency arises remains poorly understood. In this work, we first identify re-explanation-induced context perturbation as the direct cause of score variation. We then introduce a latent signal assignment hypothesis to explain why only some edges are sensitive to this perturbation, and analyze how conciseness regularization affects latent signal assignment. Given that self-inconsistent edges do not provide stable evidence for the model's prediction, we propose Self-Denoising (SD), a model-agnostic and training-free post-processing strategy that calibrates explanations with only one additional forward pass. Experiments across representative SI-GNN frameworks, backbone architectures, and benchmark datasets support our hypothesis and show that SD consistently improves explanation quality while adding only about 4--6\% computational overhead in practice.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。