Weak Robust Compatibility Between Learning Algorithms and Counterfactual Explanation Generation Algorithms

TL;DR

This paper introduces a new perspective on evaluating the robustness of counterfactual explanation algorithms by focusing on explanation strength and proposes a method called WRC-Test to generate more robust counterfactuals.

Contribution

It proposes the concept of Weak Robust Compatibility, a new robustness definition considering explanation strength, and introduces WRC-Test for improved counterfactual robustness evaluation.

Findings

WRC-Test effectively generates more robust counterfactuals.

Theoretical analysis establishes conditions for PAC WRC-Approximability.

Experiments verify the practical effectiveness of the proposed method.

Abstract

Counterfactual explanation generation is a powerful method for Explainable Artificial Intelligence. It can help users understand why machine learning models make specific decisions, and how to change those decisions. Evaluating the robustness of counterfactual explanation algorithms is therefore crucial. Previous literature has widely studied the robustness based on the perturbation of input instances. However, the robustness defined from the perspective of perturbed instances is sometimes biased, because this definition ignores the impact of learning algorithms on robustness. In this paper, we propose a more reasonable definition, Weak Robust Compatibility, based on the perspective of explanation strength. In practice, we propose WRC-Test to help us generate more robust counterfactuals. Meanwhile, we designed experiments to verify the effectiveness of WRC-Test. Theoretically, we introduce the concepts of PAC learning theory and define the concept of PAC WRC-Approximability. Based on reasonable assumptions, we establish oracle inequalities about weak robustness, which gives a sufficient condition for PAC WRC-Approximability.