MACE: An Efficient Model-Agnostic Framework for Counterfactual Explanation
Abstract
Counterfactual explanation is an important Explainable AI technique to explain machine learning predictions. Despite being studied actively, existing optimization-based methods often assume that the underlying machine-learning model is differentiable and treat categorical attributes as continuous ones, which restricts their real-world applications when categorical attributes have many different values or the model is non-differentiable. To make counterfactual explanation suitable for real-world applications, we propose a novel framework of Model-Agnostic Counterfactual Explanation (MACE), which adopts a newly designed pipeline that can efficiently handle non-differentiable machine-learning models on a large number of feature values. in our MACE approach, we propose a novel RL-based method for finding good counterfactual examples and a gradient-less descent method for improving proximity. Experiments on public datasets validate the effectiveness with better validity, sparsity and proximity.
Cite
@article{arxiv.2205.15540,
title = {MACE: An Efficient Model-Agnostic Framework for Counterfactual Explanation},
author = {Wenzhuo Yang and Jia Li and Caiming Xiong and Steven C. H. Hoi},
journal= {arXiv preprint arXiv:2205.15540},
year = {2022}
}
Comments
9 pages, 2 figures