We develop a Reinforcement Learning (RL) framework for improving an existing behavior policy via sparse, user-interpretable changes. Our goal is to make minimal changes while gaining as much benefit as possible. We define a minimal change as having a sparse, global contrastive explanation between the original and proposed policy. We improve the current policy with the constraint of keeping that global contrastive explanation short. We demonstrate our framework with a discrete MDP and a continuous 2D navigation domain.
@article{arxiv.2207.06269,
title = {Policy Optimization with Sparse Global Contrastive Explanations},
author = {Jiayu Yao and Sonali Parbhoo and Weiwei Pan and Finale Doshi-Velez},
journal= {arXiv preprint arXiv:2207.06269},
year = {2022}
}