Game-theoretic Counterfactual Explanation for Graph Neural Networks

TL;DR

The paper addresses the interpretability challenge of Graph Neural Networks (GNNs) by proposing a semivalue-based, non-learning method to generate counterfactual explanations for node classification, avoiding any additional training. It argues that Banzhaf values offer lower sample complexity and can achieve significant speedups (up to about 4x) over Shapley-based approaches, with a thresholding technique improving robustness. The method aims to provide efficient, faithful explanations by focusing on the neighborhood of the target node and demonstrates robustness and efficiency gains across multiple graph datasets. Overall, the work presents a practical, training-free pathway to interpretable GNN decisions with scalable explanation generation and competitive fidelity.

Abstract

Graph Neural Networks (GNNs) have been a powerful tool for node classification tasks in complex networks. However, their decision-making processes remain a black-box to users, making it challenging to understand the reasoning behind their predictions. Counterfactual explanations (CFE) have shown promise in enhancing the interpretability of machine learning models. Prior approaches to compute CFE for GNNS often are learning-based approaches that require training additional graphs. In this paper, we propose a semivalue-based, non-learning approach to generate CFE for node classification tasks, eliminating the need for any additional training. Our results reveals that computing Banzhaf values requires lower sample complexity in identifying the counterfactual explanations compared to other popular methods such as computing Shapley values. Our empirical evidence indicates computing Banzhaf values can achieve up to a fourfold speed up compared to Shapley values. We also design a thresholding method for computing Banzhaf values and show theoretical and empirical results on its robustness in noisy environments, making it superior to Shapley values. Furthermore, the thresholded Banzhaf values are shown to enhance efficiency without compromising the quality (i.e., fidelity) in the explanations in three popular graph datasets.