ATEX-CF: Attack-Informed Counterfactual Explanations for Graph Neural Networks

TL;DR

ATEX-CF introduces a unified framework combining adversarial attack techniques with counterfactual explanations to generate faithful, sparse, and plausible explanations for graph neural networks by jointly optimizing multiple criteria.

Contribution

This work presents the first method to integrate adversarial attack insights into counterfactual explanation generation for GNNs, allowing both edge additions and deletions.

Findings

ATEX-CF produces more faithful explanations than traditional methods.

The method achieves high plausibility and sparsity in explanations.

Experimental results on benchmarks validate its effectiveness.

Abstract

Counterfactual explanations offer an intuitive way to interpret graph neural networks (GNNs) by identifying minimal changes that alter a model's prediction, thereby answering "what must differ for a different outcome?". In this work, we propose a novel framework, ATEX-CF that unifies adversarial attack techniques with counterfactual explanation generation-a connection made feasible by their shared goal of flipping a node's prediction, yet differing in perturbation strategy: adversarial attacks often rely on edge additions, while counterfactual methods typically use deletions. Unlike traditional approaches that treat explanation and attack separately, our method efficiently integrates both edge additions and deletions, grounded in theory, leveraging adversarial insights to explore impactful counterfactuals. In addition, by jointly optimizing fidelity, sparsity, and plausibility under a constrained perturbation budget, our method produces instance-level explanations that are both informative and realistic. Experiments on synthetic and real-world node classification benchmarks demonstrate that ATEX-CF generates faithful, concise, and plausible explanations, highlighting the effectiveness of integrating adversarial insights into counterfactual reasoning for GNNs.