BriefGPT.xyz
Nov, 2024
基于模态和功能ANOVA的贝叶斯机器学习模型解释
A Bayesian explanation of machine learning models based on modes and functional ANOVA
HTML
PDF
Quan Long
TL;DR
本研究针对可解释人工智能（XAI）中的逆解释问题，提出了一种新的贝叶斯方法来识别和排名影响标签偏差的特征。与传统基于均值的方法相比，该方法更直观且稳健，能够有效解释标签值的偏差，为理解模型决策提供了新的视角。
Abstract
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 deviation of a label, find the reas
→