On the Robustness of Post-hoc GNN Explainers to Label Noise

TL;DR

This paper systematically investigates how post-hoc GNN explainers are affected by label noise, revealing their vulnerability and the impact of even minor label perturbations on explanation quality.

Contribution

It provides the first empirical analysis of the robustness of post-hoc GNN explainers against label noise, highlighting their susceptibility and effects on explanation quality.

Findings

Post-hoc GNN explainers are sensitive to label noise.

Minor label noise can significantly degrade explanation quality.

Explanation effectiveness can partially recover with increasing noise levels.

Abstract

Proposed as a solution to the inherent black-box limitations of graph neural networks (GNNs), post-hoc GNN explainers aim to provide precise and insightful explanations of the behaviours exhibited by trained GNNs. Despite their recent notable advancements in academic and industrial contexts, the robustness of post-hoc GNN explainers remains unexplored when confronted with label noise. To bridge this gap, we conduct a systematic empirical investigation to evaluate the efficacy of diverse post-hoc GNN explainers under varying degrees of label noise. Our results reveal several key insights: Firstly, post-hoc GNN explainers are susceptible to label perturbations. Secondly, even minor levels of label noise, inconsequential to GNN performance, harm the quality of generated explanations substantially. Lastly, we engage in a discourse regarding the progressive recovery of explanation effectiveness with escalating noise levels.