ExplainabilityTransparencyAccountabilitycs.LG
BEExAI: Benchmark to Evaluate Explainable AI
Samuel Sithakoul, Sara Meftah, Clément Feutry
July 29, 2024
xAI
20 citations
3 influential
Key Takeaways
-
There is no standard way to evaluate explanation quality: The paper identifies that while many methods exist to explain AI model decisions, there is currently "a lack of cohesive approach and consensus on the methodology" for measuring whether these explanations are actually good or reliable.
-
Evaluation of explanations is becoming urgent: As AI models grow more complex and are deployed across diverse applications, the paper emphasizes that "the need for a reliable way of measuring the quality and correctness of explanations is becoming critical" for responsible AI deployment.
-
BEExAI provides a benchmarking solution: The paper proposes BEExAI as a tool that enables "large-scale comparison of different post-hoc XAI methods" using standardized evaluation metrics, offering a practical framework for organizations to assess which explanation methods work best.
-
Post-hoc explanations dominate but need evaluation: The paper notes that "numerous post-hoc attribution methods" have been developed to help understand black-box AI model outputs, but without proper evaluation frameworks, it's difficult to know which methods actually provide trustworthy explanations.
Abstract
Recent research in explainability has given rise to numerous post-hoc attribution methods aimed at enhancing our comprehension of the outputs of black-box machine learning models. However, evaluating the quality of explanations lacks a cohesive approach and a consensus on the methodology for deriving quantitative metrics that gauge the efficacy of explainability post-hoc attribution methods. Furthermore, with the development of increasingly complex deep learning models for diverse data applications, the need for a reliable way of measuring the quality and correctness of explanations is becoming critical. We address this by proposing BEExAI, a benchmark tool that allows large-scale comparison of different post-hoc XAI methods, employing a set of selected evaluation metrics.