Redefining Counterfactual Explanations for Reinforcement Learning: Overview, Challenges and Opportunities

TL;DR

This paper reviews counterfactual explanations in supervised learning, identifies challenges in applying them to reinforcement learning, and proposes a redefinition and future research directions for their use in RL systems.

Contribution

It redefines counterfactual explanations specifically for reinforcement learning and outlines key challenges and opportunities for their development and application.

Findings

Counterfactual explanations are well-studied in supervised learning.

Applying counterfactuals to RL faces unique challenges.

The paper proposes a new framework for RL counterfactuals.

Abstract

While AI algorithms have shown remarkable success in various fields, their lack of transparency hinders their application to real-life tasks. Although explanations targeted at non-experts are necessary for user trust and human-AI collaboration, the majority of explanation methods for AI are focused on developers and expert users. Counterfactual explanations are local explanations that offer users advice on what can be changed in the input for the output of the black-box model to change. Counterfactuals are user-friendly and provide actionable advice for achieving the desired output from the AI system. While extensively researched in supervised learning, there are few methods applying them to reinforcement learning (RL). In this work, we explore the reasons for the underrepresentation of a powerful explanation method in RL. We start by reviewing the current work in counterfactual explanations in supervised learning. Additionally, we explore the differences between counterfactual explanations in supervised learning and RL and identify the main challenges that prevent the adoption of methods from supervised in reinforcement learning. Finally, we redefine counterfactuals for RL and propose research directions for implementing counterfactuals in RL.