The Generalizability of Explanations
This work addresses the challenge of evaluating explainability methods for machine learning practitioners, but it is incremental as it builds on existing evaluation categories.
The paper tackles the problem of objectively evaluating explainability methods by proposing a novel evaluation methodology based on generalizability, using an Autoencoder to assess learnability and plausibility, and finds that SmoothGrad smoothing significantly enhances generalizability.
Due to the absence of ground truth, objective evaluation of explainability methods is an essential research direction. So far, the vast majority of evaluations can be summarized into three categories, namely human evaluation, sensitivity testing, and salinity check. This work proposes a novel evaluation methodology from the perspective of generalizability. We employ an Autoencoder to learn the distributions of the generated explanations and observe their learnability as well as the plausibility of the learned distributional features. We first briefly demonstrate the evaluation idea of the proposed approach at LIME, and then quantitatively evaluate multiple popular explainability methods. We also find that smoothing the explanations with SmoothGrad can significantly enhance the generalizability of explanations.