TY - UNPB
T1 - From Anecdotal Evidence to Quantitative Evaluation Methods
T2 - A Systematic Review on Evaluating Explainable AI
AU - Nauta, Meike
AU - Trienes, Jan
AU - Pathak, Shreyasi
AU - Nguyen, Elisa
AU - Peters, Michelle
AU - Schmitt, Yasmin
AU - Schlötterer, Jörg
AU - van Keulen, Maurice
AU - Seifert, Christin
PY - 2022/5/31
Y1 - 2022/5/31
N2 - The rising popularity of explainable artificial intelligence (XAI) to understand high-performing black boxes, also 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 practice 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. We also contribute to the call for objective, quantifiable evaluation methods by presenting an extensive overview of quantitative XAI evaluation methods. This systematic collection of evaluation methods provides researchers and practitioners with concrete tools to thoroughly validate, benchmark and compare new and existing XAI methods. This also opens up opportunities to include quantitative metrics as optimization criteria during model training in order to optimize for accuracy and interpretability simultaneously.
AB - The rising popularity of explainable artificial intelligence (XAI) to understand high-performing black boxes, also 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 practice 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. We also contribute to the call for objective, quantifiable evaluation methods by presenting an extensive overview of quantitative XAI evaluation methods. This systematic collection of evaluation methods provides researchers and practitioners with concrete tools to thoroughly validate, benchmark and compare new and existing XAI methods. This also opens up opportunities to include quantitative metrics as optimization criteria during model training in order to optimize for accuracy and interpretability simultaneously.
U2 - 10.48550/arXiv.2201.08164
DO - 10.48550/arXiv.2201.08164
M3 - Preprint
BT - From Anecdotal Evidence to Quantitative Evaluation Methods
PB - ArXiv.org
ER -