Objective Metrics for Human-Subjects Evaluation in Explainable Reinforcement Learning

TL;DR

This paper advocates for using objective, observable human metrics in explainable reinforcement learning to improve evaluation reproducibility and comparability, moving beyond subjective assessments.

Contribution

It introduces and compares objective evaluation methodologies for explanations in XRL, emphasizing observable human behavior for debugging and teaming tasks.

Findings

Objective metrics enable more reproducible evaluations.

Standardized benchmarks facilitate better comparison across studies.

Objective and subjective metrics complement each other for comprehensive validation.

Abstract

Explanation is a fundamentally human process. Understanding the goal and audience of the explanation is vital, yet existing work on explainable reinforcement learning (XRL) routinely does not consult humans in their evaluations. Even when they do, they routinely resort to subjective metrics, such as confidence or understanding, that can only inform researchers of users' opinions, not their practical effectiveness for a given problem. This paper calls on researchers to use objective human metrics for explanation evaluations based on observable and actionable behaviour to build more reproducible, comparable, and epistemically grounded research. To this end, we curate, describe, and compare several objective evaluation methodologies for applying explanations to debugging agent behaviour and supporting human-agent teaming, illustrating our proposed methods using a novel grid-based environment. We discuss how subjective and objective metrics complement each other to provide holistic validation and how future work needs to utilise standardised benchmarks for testing to enable greater comparisons between research.