{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/data-efficient-off-policy-policy-evaluation","title":"Data-Efficient Off-Policy Policy Evaluation for Reinforcement Learning","arxiv_id":"1604.00923","date":"2016-04-04","proceeding":null,"authors":["Philip S. Thomas","Emma Brunskill"],"abstract":"In this paper we present a new way of predicting the performance of a\nreinforcement learning policy given historical data that may have been\ngenerated by a different policy. The ability to evaluate a policy from\nhistorical data is important for applications where the deployment of a bad\npolicy can be dangerous or costly. We show empirically that our algorithm\nproduces estimates that often have orders of magnitude lower mean squared error\nthan existing methods---it makes more efficient use of the available data. Our\nnew estimator is based on two advances: an extension of the doubly robust\nestimator (Jiang and Li, 2015), and a new way to mix between model based\nestimates and importance sampling based estimates.","url_abs":"http://arxiv.org/abs/1604.00923v1","url_pdf":"http://arxiv.org/pdf/1604.00923v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"data-efficient-off-policy-policy-evaluation","repo_url":"https://github.com/ShawnBLYU/offline_rl_envs","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"data-efficient-off-policy-policy-evaluation","repo_url":"https://github.com/facebookresearch/Horizon","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"data-efficient-off-policy-policy-evaluation","repo_url":"https://github.com/facebookresearch/ReAgent","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"reinforcement-learning-2","task_name":"reinforcement-learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1604.00923","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}