{"id":453927,"date":"2018-01-12T06:22:59","date_gmt":"2018-01-12T14:22:59","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-research-item&#038;p=453927"},"modified":"2018-10-16T20:10:29","modified_gmt":"2018-10-17T03:10:29","slug":"safe-policy-improvement-baseline-bootstrapping","status":"publish","type":"msr-research-item","link":"https:\/\/www.microsoft.com\/en-us\/research\/publication\/safe-policy-improvement-baseline-bootstrapping\/","title":{"rendered":"Safe Policy Improvement with Baseline Bootstrapping"},"content":{"rendered":"<p><span style=\"color: #000000;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: 14.4px;font-style: normal;font-weight: 400;letter-spacing: normal;text-align: start;text-indent: 0px;background-color: #ffffff;float: none\">A common goal in Reinforcement Learning is to derive a good strategy given a limited batch of data. In this paper, we adopt the safe policy improvement (SPI) approach: we compute a target policy guaranteed to perform at least as well as a given baseline policy. Our SPI strategy, inspired by the knows-what-it-knows paradigms, consists in bootstrapping the target policy with the baseline policy when it does not know. We develop two computationally efficient bootstrapping algorithms, a value-based and a policy-based, both accompanied with theoretical SPI bounds. Three algorithm variants are proposed. We empirically show the literature algorithms limits on a small stochastic gridworld problem, and then demonstrate that our five algorithms not only improve the worst case scenarios, but also the mean performance.<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>A common goal in Reinforcement Learning is to derive a good strategy given a limited batch of data. In this paper, we adopt the safe policy improvement (SPI) approach: we compute a target policy guaranteed to perform at least as well as a given baseline policy. Our SPI strategy, inspired by the knows-what-it-knows paradigms, consists [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"footnotes":""},"msr-content-type":[3],"msr-research-highlight":[],"research-area":[13556],"msr-publication-type":[193715],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-locale":[268875],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-453927","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-locale-en_us"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2017-12-20","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"8+1+8","msr_chapter":"","msr_isbn":"","msr_journal":"arXiv preprint arXiv:1712.06924","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"http:\/\/YToxOntpOjA7YToxOntzOjM6InVybCI7czozNjoiaHR0cHM6Ly9hcnhpdi5vcmcvcGRmLzE3MTIuMDY5MjQucGRmIjt9fQ==","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"453939","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"file","title":"SPIBB","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2018\/01\/SPIBB.pdf","id":453939,"label_id":0}],"msr_related_uploader":"","msr_attachments":[],"msr-author-ordering":[{"type":"user_nicename","value":"rolaroch","user_id":36623,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=rolaroch"},{"type":"text","value":"Paul Trichelair","user_id":0,"rest_url":false}],"msr_impact_theme":[],"msr_research_lab":[437514],"msr_event":[],"msr_group":[395930,663258],"msr_project":[],"publication":[],"video":[],"download":[],"msr_publication_type":"article","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/453927"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":1,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/453927\/revisions"}],"predecessor-version":[{"id":453930,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/453927\/revisions\/453930"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=453927"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=453927"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=453927"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=453927"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=453927"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=453927"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=453927"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=453927"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=453927"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=453927"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=453927"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=453927"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=453927"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=453927"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=453927"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}