From Anecdotal Evidence to Quantitative Evaluation Methods: A Systematic Review on Evaluating Explainable AI

TL;DR

This paper systematically reviews evaluation practices for explainable AI, introducing a comprehensive categorization scheme and highlighting the need for objective, quantitative assessment methods to improve explanation quality.

Contribution

It presents the Co-12 properties as a framework for evaluating XAI explanations and compiles a wide range of quantitative evaluation methods from recent literature.

Findings

1 in 3 papers rely on anecdotal evidence for evaluation

1 in 5 papers involve user-based evaluation methods

The Co-12 scheme enables systematic assessment of explanation quality

Abstract

The rising popularity of explainable artificial intelligence (XAI) to understand high-performing black boxes raised the question of how to evaluate explanations of machine learning (ML) models. While interpretability and explainability are often presented as a subjectively validated binary property, we consider it a multi-faceted concept. We identify 12 conceptual properties, such as Compactness and Correctness, that should be evaluated for comprehensively assessing the quality of an explanation. Our so-called Co-12 properties serve as categorization scheme for systematically reviewing the evaluation practices of more than 300 papers published in the last 7 years at major AI and ML conferences that introduce an XAI method. We find that 1 in 3 papers evaluate exclusively with anecdotal evidence, and 1 in 5 papers evaluate with users. This survey also contributes to the call for objective, quantifiable evaluation methods by presenting an extensive overview of quantitative XAI evaluation methods. Our systematic collection of evaluation methods provides researchers and practitioners with concrete tools to thoroughly validate, benchmark and compare new and existing XAI methods. The Co-12 categorization scheme and our identified evaluation methods open up opportunities to include quantitative metrics as optimization criteria during model training in order to optimize for accuracy and interpretability simultaneously.