Counterfactual Explanations for Continuous Action Reinforcement Learning

TL;DR

This paper introduces a new method for generating counterfactual explanations in continuous action reinforcement learning, enhancing interpretability and trustworthiness in domains like healthcare and robotics.

Contribution

It presents a novel approach that computes alternative action sequences for continuous RL, considering constraints and improving interpretability over prior methods.

Findings

Effective in Diabetes Control and Lunar Lander domains

Demonstrates improved interpretability and trustworthiness

Efficient computation of counterfactuals for continuous actions

Abstract

Reinforcement Learning (RL) has shown great promise in domains like healthcare and robotics but often struggles with adoption due to its lack of interpretability. Counterfactual explanations, which address "what if" scenarios, provide a promising avenue for understanding RL decisions but remain underexplored for continuous action spaces. We propose a novel approach for generating counterfactual explanations in continuous action RL by computing alternative action sequences that improve outcomes while minimizing deviations from the original sequence. Our approach leverages a distance metric for continuous actions and accounts for constraints such as adhering to predefined policies in specific states. Evaluations in two RL domains, Diabetes Control and Lunar Lander, demonstrate the effectiveness, efficiency, and generalization of our approach, enabling more interpretable and trustworthy RL applications.