On the stability, correctness and plausibility of visual explanation methods based on feature importance
This work addresses the challenge of reliable evaluation in Explainable AI for researchers and practitioners, but it is incremental as it builds on existing metrics and focuses on specific properties.
The paper tackles the problem of evaluating visual explanation methods for image classifiers by analyzing the alignment between stability, correctness, and plausibility metrics, showing that existing metrics often disagree and proposing new metrics for stability and correctness that account for local model behavior.
In the field of Explainable AI, multiples evaluation metrics have been proposed in order to assess the quality of explanation methods w.r.t. a set of desired properties. In this work, we study the articulation between the stability, correctness and plausibility of explanations based on feature importance for image classifiers. We show that the existing metrics for evaluating these properties do not always agree, raising the issue of what constitutes a good evaluation metric for explanations. Finally, in the particular case of stability and correctness, we show the possible limitations of some evaluation metrics and propose new ones that take into account the local behaviour of the model under test.