Exploring Counterfactual Explanations Through the Lens of Adversarial Examples: A Theoretical and Empirical Analysis

TL;DR

This paper investigates the relationship between counterfactual explanations and adversarial examples in machine learning, providing theoretical bounds and empirical validation to deepen understanding of their similarities and implications.

Contribution

It formalizes the connection between counterfactual explanations and adversarial examples, deriving bounds and analyzing their equivalence conditions.

Findings

Counterfactual explanations and adversarial examples are closely related under certain conditions.

Theoretical upper bounds on the distance between counterfactuals and adversarial examples.

Empirical validation on real-world datasets supports the theoretical analysis.

Abstract

As machine learning (ML) models become more widely deployed in high-stakes applications, counterfactual explanations have emerged as key tools for providing actionable model explanations in practice. Despite the growing popularity of counterfactual explanations, a deeper understanding of these explanations is still lacking. In this work, we systematically analyze counterfactual explanations through the lens of adversarial examples. We do so by formalizing the similarities between popular counterfactual explanation and adversarial example generation methods identifying conditions when they are equivalent. We then derive the upper bounds on the distances between the solutions output by counterfactual explanation and adversarial example generation methods, which we validate on several real-world data sets. By establishing these theoretical and empirical similarities between counterfactual explanations and adversarial examples, our work raises fundamental questions about the design and development of existing counterfactual explanation algorithms.