High-Confidence Off-Policy Evaluation
Abstract
Many reinforcement learning algorithms use trajectories collected from the execution of one or more policies to propose a new policy. Because execution of a bad policy can be costly or dangerous, techniques for evaluating the performance of the new policy without requiring its execution have been of recent interest in industry. Such off-policy evaluation methods, which estimate the performance of a policy using trajectories collected from the execution of other policies, heretofore have not provided confidences regarding the accuracy of their estimates. In this paper we propose an off-policy method for computing a lower confidence bound on the expected return of a policy.
Cite
Text
Thomas et al. "High-Confidence Off-Policy Evaluation." AAAI Conference on Artificial Intelligence, 2015. doi:10.1609/AAAI.V29I1.9541Markdown
[Thomas et al. "High-Confidence Off-Policy Evaluation." AAAI Conference on Artificial Intelligence, 2015.](https://mlanthology.org/aaai/2015/thomas2015aaai-high/) doi:10.1609/AAAI.V29I1.9541BibTeX
@inproceedings{thomas2015aaai-high,
title = {{High-Confidence Off-Policy Evaluation}},
author = {Thomas, Philip S. and Theocharous, Georgios and Ghavamzadeh, Mohammad},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2015},
pages = {3000-3006},
doi = {10.1609/AAAI.V29I1.9541},
url = {https://mlanthology.org/aaai/2015/thomas2015aaai-high/}
}