Price of Safety in Linear Best Arm Identification

Price of Safety in Linear Best Arm Identification

We introduce the safe best-arm identification framework with linear feedback, where the agent is subject to some stage-wise safety constraint that linearly depends on an unknown parameter vector. The agent must take actions in a conservative way so as to ensure that the safety constraint is not viol...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Shang, Xuedong, Colin, Igor, Barlier, Merwan, Cherkaoui, Hamza
Format: Artikel
Sprache:eng
Schlagworte:
Computer Science - Learning
Statistics - Machine Learning
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Shang, Xuedong
Colin, Igor
Barlier, Merwan
Cherkaoui, Hamza
description We introduce the safe best-arm identification framework with linear feedback, where the agent is subject to some stage-wise safety constraint that linearly depends on an unknown parameter vector. The agent must take actions in a conservative way so as to ensure that the safety constraint is not violated with high probability at each round. Ways of leveraging the linear structure for ensuring safety has been studied for regret minimization, but not for best-arm identification to the best our knowledge. We propose a gap-based algorithm that achieves meaningful sample complexity while ensuring the stage-wise safety. We show that we pay an extra term in the sample complexity due to the forced exploration phase incurred by the additional safety constraint. Experimental illustrations are provided to justify the design of our algorithm.
doi_str_mv 10.48550/arxiv.2309.08709
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2309_08709</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2309_08709</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-588a02ddb3d0f2f21e3d0ac49307f5f56f1a13f8f85d1b9560d5a87d7838a8e83</originalsourceid><addsrcrecordid>eNotzs2KwjAUhuFsXEj1AlyZG2jnpDHNydIRnREKM6D7ctrkQECrxCJ69_7MrL539fEIMVNQLNAY-KB0i9ei1OAKQAtuLNRvil2QJ5Y74jDcZexlHftASX6GyyCX6Si3PvRD5NjREE_9RIyYDpcw_d9M7Dfr_eo7r3--tqtlnVNlXW4QCUrvW-2BSy5VeAZ1C6fBsmFTsSKlGRmNV60zFXhDaL1FjYQBdSbmf7dvc3NO8Ujp3rzszduuH_WGPXc</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Price of Safety in Linear Best Arm Identification</title><source>arXiv.org</source><creator>Shang, Xuedong ; Colin, Igor ; Barlier, Merwan ; Cherkaoui, Hamza</creator><creatorcontrib>Shang, Xuedong ; Colin, Igor ; Barlier, Merwan ; Cherkaoui, Hamza</creatorcontrib><description>We introduce the safe best-arm identification framework with linear feedback, where the agent is subject to some stage-wise safety constraint that linearly depends on an unknown parameter vector. The agent must take actions in a conservative way so as to ensure that the safety constraint is not violated with high probability at each round. Ways of leveraging the linear structure for ensuring safety has been studied for regret minimization, but not for best-arm identification to the best our knowledge. We propose a gap-based algorithm that achieves meaningful sample complexity while ensuring the stage-wise safety. We show that we pay an extra term in the sample complexity due to the forced exploration phase incurred by the additional safety constraint. Experimental illustrations are provided to justify the design of our algorithm.</description><identifier>DOI: 10.48550/arxiv.2309.08709</identifier><language>eng</language><subject>Computer Science - Learning ; Statistics - Machine Learning</subject><creationdate>2023-09</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2309.08709$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2309.08709$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Shang, Xuedong</creatorcontrib><creatorcontrib>Colin, Igor</creatorcontrib><creatorcontrib>Barlier, Merwan</creatorcontrib><creatorcontrib>Cherkaoui, Hamza</creatorcontrib><title>Price of Safety in Linear Best Arm Identification</title><description>We introduce the safe best-arm identification framework with linear feedback, where the agent is subject to some stage-wise safety constraint that linearly depends on an unknown parameter vector. The agent must take actions in a conservative way so as to ensure that the safety constraint is not violated with high probability at each round. Ways of leveraging the linear structure for ensuring safety has been studied for regret minimization, but not for best-arm identification to the best our knowledge. We propose a gap-based algorithm that achieves meaningful sample complexity while ensuring the stage-wise safety. We show that we pay an extra term in the sample complexity due to the forced exploration phase incurred by the additional safety constraint. Experimental illustrations are provided to justify the design of our algorithm.</description><subject>Computer Science - Learning</subject><subject>Statistics - Machine Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotzs2KwjAUhuFsXEj1AlyZG2jnpDHNydIRnREKM6D7ctrkQECrxCJ69_7MrL539fEIMVNQLNAY-KB0i9ei1OAKQAtuLNRvil2QJ5Y74jDcZexlHftASX6GyyCX6Si3PvRD5NjREE_9RIyYDpcw_d9M7Dfr_eo7r3--tqtlnVNlXW4QCUrvW-2BSy5VeAZ1C6fBsmFTsSKlGRmNV60zFXhDaL1FjYQBdSbmf7dvc3NO8Ujp3rzszduuH_WGPXc</recordid><startdate>20230915</startdate><enddate>20230915</enddate><creator>Shang, Xuedong</creator><creator>Colin, Igor</creator><creator>Barlier, Merwan</creator><creator>Cherkaoui, Hamza</creator><scope>AKY</scope><scope>EPD</scope><scope>GOX</scope></search><sort><creationdate>20230915</creationdate><title>Price of Safety in Linear Best Arm Identification</title><author>Shang, Xuedong ; Colin, Igor ; Barlier, Merwan ; Cherkaoui, Hamza</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-588a02ddb3d0f2f21e3d0ac49307f5f56f1a13f8f85d1b9560d5a87d7838a8e83</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Learning</topic><topic>Statistics - Machine Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Shang, Xuedong</creatorcontrib><creatorcontrib>Colin, Igor</creatorcontrib><creatorcontrib>Barlier, Merwan</creatorcontrib><creatorcontrib>Cherkaoui, Hamza</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv Statistics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Shang, Xuedong</au><au>Colin, Igor</au><au>Barlier, Merwan</au><au>Cherkaoui, Hamza</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Price of Safety in Linear Best Arm Identification</atitle><date>2023-09-15</date><risdate>2023</risdate><abstract>We introduce the safe best-arm identification framework with linear feedback, where the agent is subject to some stage-wise safety constraint that linearly depends on an unknown parameter vector. The agent must take actions in a conservative way so as to ensure that the safety constraint is not violated with high probability at each round. Ways of leveraging the linear structure for ensuring safety has been studied for regret minimization, but not for best-arm identification to the best our knowledge. We propose a gap-based algorithm that achieves meaningful sample complexity while ensuring the stage-wise safety. We show that we pay an extra term in the sample complexity due to the forced exploration phase incurred by the additional safety constraint. Experimental illustrations are provided to justify the design of our algorithm.</abstract><doi>10.48550/arxiv.2309.08709</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2309.08709
ispartof
issn
language eng
recordid cdi_arxiv_primary_2309_08709
source arXiv.org
subjects Computer Science - Learning
Statistics - Machine Learning
title Price of Safety in Linear Best Arm Identification
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-01T07%3A08%3A08IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Price%20of%20Safety%20in%20Linear%20Best%20Arm%20Identification&rft.au=Shang,%20Xuedong&rft.date=2023-09-15&rft_id=info:doi/10.48550/arxiv.2309.08709&rft_dat=%3Carxiv_GOX%3E2309_08709%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true