arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.07955cs.LGstat.ML
收藏

可解释图神经网络的可解释性如何?

How Interpretable Are Interpretable Graph Neural Networks?

Tencent AI Lab(腾讯人工智能实验室);The Chinese University of Hong Kong(香港中文大学)

Yongqiang Chen, Yatao Bian, Bo Han, James Cheng

AI总结:

本文提出 SubMT 理论框架分析可解释图神经网络的子图学习局限,并设计 GMT 架构以更准确近似 SubMT,在 12 个基准上显著提升可解释性与泛化性。

摘要:

可解释图神经网络(XGNNs)被广泛应用于涉及图结构数据的各类科学应用中。现有 XGNNs 主要采用基于注意力的机制来学习边或节点的重要性,以提取可解释子图并利用其进行预测。然而,这些方法的表示特性与局限性仍未得到充分研究。在本工作中,我们提出一个理论框架,利用子图分布的多线性扩展来形式化可解释子图学习,并将其命名为子图多线性扩展(SubMT)。提取期望的可解释子图需要对 SubMT 进行准确近似,但我们发现现有 XGNNs 在拟合 SubMT 时可能存在巨大差距。因此,SubMT 近似失败会导致所提取子图的可解释性退化。为缓解这一问题,我们设计了一种名为 Graph Multilinear neT(GMT)的新型 XGNN 架构,该架构在近似 SubMT 方面被证明更强大。我们在多个图分类基准上通过实验验证了理论发现。结果表明,在 12 个常规图和几何图基准上,GMT 在可解释性和泛化能力两方面均优于最先进方法,提升幅度最高达 10%。

英文摘要:

Interpretable graph neural networks (XGNNs ) are widely adopted in various scientific applications involving graph-structured data. Existing XGNNs predominantly adopt the attention-based mechanism to learn edge or node importance for extracting and making predictions with the interpretable subgraph. However, the representational properties and limitations of these methods remain inadequately explored. In this work, we present a theoretical framework that formulates interpretable subgraph learning with the multilinear extension of the subgraph distribution, coined as subgraph multilinear extension (SubMT). Extracting the desired interpretable subgraph requires an accurate approximation of SubMT, yet we find that the existing XGNNs can have a huge gap in fitting SubMT. Consequently, the SubMT approximation failure will lead to the degenerated interpretability of the extracted subgraphs. To mitigate the issue, we design a new XGNN architecture called Graph Multilinear neT (GMT), which is provably more powerful in approximating SubMT. We empirically validate our theoretical findings on a number of graph classification benchmarks. The results demonstrate that GMT outperforms the state-of-the-art up to 10% in terms of both interpretability and generalizability across 12 regular and geometric graph benchmarks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑