Hypothesis Transfer in Bandits by Weighted Models

We consider the problem of contextual multi-armed bandits in the setting of hypothesis transfer learning. That is, we assume having access to a previously learned model on an unobserved set of contexts, and we leverage it in order to accelerate exploration on a new bandit problem. Our transfer strat...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Bilaj, Steven, Dhouib, Sofien, Maghsudi, Setareh
Format: Artikel
Sprache:eng
Schlagworte:
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 Bilaj, Steven
Dhouib, Sofien
Maghsudi, Setareh
description We consider the problem of contextual multi-armed bandits in the setting of hypothesis transfer learning. That is, we assume having access to a previously learned model on an unobserved set of contexts, and we leverage it in order to accelerate exploration on a new bandit problem. Our transfer strategy is based on a re-weighting scheme for which we show a reduction in the regret over the classic Linear UCB when transfer is desired, while recovering the classic regret rate when the two tasks are unrelated. We further extend this method to an arbitrary amount of source models, where the algorithm decides which model is preferred at each time step. Additionally we discuss an approach where a dynamic convex combination of source models is given in terms of a biased regularization term in the classic LinUCB algorithm. The algorithms and the theoretical analysis of our proposed methods substantiated by empirical evaluations on simulated and real-world data.
doi_str_mv 10.48550/arxiv.2211.07387
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2211_07387</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2211_07387</sourcerecordid><originalsourceid>FETCH-LOGICAL-a677-7db3e012e1a2a96d5e4812b2488feb5ada60535ce1fabcc23f72aa738c9e78b13</originalsourceid><addsrcrecordid>eNotzr1uwjAUQGEvDBX0ATrhF0jqa8exGSlqCxKoS6SO0XV8DZYgIDuqmrdv-ZnOdvQx9gKirKzW4hXTb_wppQQohVHWPDFYj5fzcKAcM28S9jlQ4rHnb9j7OGTuRv5NcX8YyPPd2dMxz9gk4DHT86NT1ny8N6t1sf363KyW2wJrYwrjnSIBkgAlLmqvqbIgnaysDeQ0eqyFVrojCOi6TqpgJOK_qVuQsQ7UlM3v25u5vaR4wjS2V3t7s6s_iDE-rA</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Hypothesis Transfer in Bandits by Weighted Models</title><source>arXiv.org</source><creator>Bilaj, Steven ; Dhouib, Sofien ; Maghsudi, Setareh</creator><creatorcontrib>Bilaj, Steven ; Dhouib, Sofien ; Maghsudi, Setareh</creatorcontrib><description>We consider the problem of contextual multi-armed bandits in the setting of hypothesis transfer learning. That is, we assume having access to a previously learned model on an unobserved set of contexts, and we leverage it in order to accelerate exploration on a new bandit problem. Our transfer strategy is based on a re-weighting scheme for which we show a reduction in the regret over the classic Linear UCB when transfer is desired, while recovering the classic regret rate when the two tasks are unrelated. We further extend this method to an arbitrary amount of source models, where the algorithm decides which model is preferred at each time step. Additionally we discuss an approach where a dynamic convex combination of source models is given in terms of a biased regularization term in the classic LinUCB algorithm. The algorithms and the theoretical analysis of our proposed methods substantiated by empirical evaluations on simulated and real-world data.</description><identifier>DOI: 10.48550/arxiv.2211.07387</identifier><language>eng</language><subject>Computer Science - Learning</subject><creationdate>2022-11</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>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2211.07387$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2211.07387$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Bilaj, Steven</creatorcontrib><creatorcontrib>Dhouib, Sofien</creatorcontrib><creatorcontrib>Maghsudi, Setareh</creatorcontrib><title>Hypothesis Transfer in Bandits by Weighted Models</title><description>We consider the problem of contextual multi-armed bandits in the setting of hypothesis transfer learning. That is, we assume having access to a previously learned model on an unobserved set of contexts, and we leverage it in order to accelerate exploration on a new bandit problem. Our transfer strategy is based on a re-weighting scheme for which we show a reduction in the regret over the classic Linear UCB when transfer is desired, while recovering the classic regret rate when the two tasks are unrelated. We further extend this method to an arbitrary amount of source models, where the algorithm decides which model is preferred at each time step. Additionally we discuss an approach where a dynamic convex combination of source models is given in terms of a biased regularization term in the classic LinUCB algorithm. The algorithms and the theoretical analysis of our proposed methods substantiated by empirical evaluations on simulated and real-world data.</description><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotzr1uwjAUQGEvDBX0ATrhF0jqa8exGSlqCxKoS6SO0XV8DZYgIDuqmrdv-ZnOdvQx9gKirKzW4hXTb_wppQQohVHWPDFYj5fzcKAcM28S9jlQ4rHnb9j7OGTuRv5NcX8YyPPd2dMxz9gk4DHT86NT1ny8N6t1sf363KyW2wJrYwrjnSIBkgAlLmqvqbIgnaysDeQ0eqyFVrojCOi6TqpgJOK_qVuQsQ7UlM3v25u5vaR4wjS2V3t7s6s_iDE-rA</recordid><startdate>20221114</startdate><enddate>20221114</enddate><creator>Bilaj, Steven</creator><creator>Dhouib, Sofien</creator><creator>Maghsudi, Setareh</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20221114</creationdate><title>Hypothesis Transfer in Bandits by Weighted Models</title><author>Bilaj, Steven ; Dhouib, Sofien ; Maghsudi, Setareh</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a677-7db3e012e1a2a96d5e4812b2488feb5ada60535ce1fabcc23f72aa738c9e78b13</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Bilaj, Steven</creatorcontrib><creatorcontrib>Dhouib, Sofien</creatorcontrib><creatorcontrib>Maghsudi, Setareh</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Bilaj, Steven</au><au>Dhouib, Sofien</au><au>Maghsudi, Setareh</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Hypothesis Transfer in Bandits by Weighted Models</atitle><date>2022-11-14</date><risdate>2022</risdate><abstract>We consider the problem of contextual multi-armed bandits in the setting of hypothesis transfer learning. That is, we assume having access to a previously learned model on an unobserved set of contexts, and we leverage it in order to accelerate exploration on a new bandit problem. Our transfer strategy is based on a re-weighting scheme for which we show a reduction in the regret over the classic Linear UCB when transfer is desired, while recovering the classic regret rate when the two tasks are unrelated. We further extend this method to an arbitrary amount of source models, where the algorithm decides which model is preferred at each time step. Additionally we discuss an approach where a dynamic convex combination of source models is given in terms of a biased regularization term in the classic LinUCB algorithm. The algorithms and the theoretical analysis of our proposed methods substantiated by empirical evaluations on simulated and real-world data.</abstract><doi>10.48550/arxiv.2211.07387</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2211.07387
ispartof
issn
language eng
recordid cdi_arxiv_primary_2211_07387
source arXiv.org
subjects Computer Science - Learning
title Hypothesis Transfer in Bandits by Weighted Models
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-06T10%3A10%3A23IST&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=Hypothesis%20Transfer%20in%20Bandits%20by%20Weighted%20Models&rft.au=Bilaj,%20Steven&rft.date=2022-11-14&rft_id=info:doi/10.48550/arxiv.2211.07387&rft_dat=%3Carxiv_GOX%3E2211_07387%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