A Bayesian explanation of machine learning models based on modes and functional ANOVA

Quan Long

Most methods in explainable AI (XAI) focus on providing reasons for the prediction of a given set of features. However, we solve an inverse explanation problem, i.e., given the dev…