<?xml version="1.0" encoding="utf-8"?>
<TEI xmlns="http://www.tei-c.org/ns/1.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:hal="http://hal.archives-ouvertes.fr/" xmlns:gml="http://www.opengis.net/gml/3.3/" xmlns:gmlce="http://www.opengis.net/gml/3.3/ce" version="1.1" xsi:schemaLocation="http://www.tei-c.org/ns/1.0 http://api.archives-ouvertes.fr/documents/aofr-sword.xsd">
  <teiHeader>
    <fileDesc>
      <titleStmt>
        <title>HAL TEI export of hal-03262006v1</title>
      </titleStmt>
      <publicationStmt>
        <distributor>CCSD</distributor>
        <availability status="restricted">
          <licence target="https://creativecommons.org/publicdomain/zero/1.0/">CC0 1.0 - Universal</licence>
        </availability>
        <date when="2026-05-25T07:32:37+02:00"/>
      </publicationStmt>
      <sourceDesc>
        <p part="N">HAL API Platform</p>
      </sourceDesc>
    </fileDesc>
  </teiHeader>
  <text>
    <body>
      <listBibl>
        <biblFull>
          <titleStmt>
            <title xml:lang="en">Reinforcement Learning for Markovian Bandits: Is Posterior Sampling more Scalable than Optimism?</title>
            <author role="aut">
              <persName>
                <forename type="first">Nicolas</forename>
                <surname>Gast</surname>
              </persName>
              <email type="md5">52c7b3fb122eece8a8132fbb2a32bc68</email>
              <email type="domain">inria.fr</email>
              <idno type="idhal" notation="string">nicolas-gast</idno>
              <idno type="idhal" notation="numeric">1247</idno>
              <idno type="halauthorid" notation="string">17248-1247</idno>
              <idno type="IDREF">https://www.idref.fr/233247874</idno>
              <idno type="ORCID">https://orcid.org/0000-0001-6884-8698</idno>
              <affiliation ref="#struct-1043327"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Bruno</forename>
                <surname>Gaujal</surname>
              </persName>
              <email type="md5">8673cdd3577f13f3e0da99ccfccbabe4</email>
              <email type="domain">imag.fr</email>
              <idno type="idhal" notation="string">bruno-gaujal</idno>
              <idno type="idhal" notation="numeric">11644</idno>
              <idno type="halauthorid" notation="string">10749-11644</idno>
              <idno type="ORCID">https://orcid.org/0000-0001-9081-8401</idno>
              <idno type="IDREF">https://www.idref.fr/074658441</idno>
              <affiliation ref="#struct-1043327"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Kimang</forename>
                <surname>Khun</surname>
              </persName>
              <email type="md5">6805820370ed96ffa0da1e84711e2284</email>
              <email type="domain">gmail.com</email>
              <idno type="idhal" notation="string">kimang-khun</idno>
              <idno type="idhal" notation="numeric">1102232</idno>
              <idno type="halauthorid" notation="string">2230586-1102232</idno>
              <idno type="ORCID">https://orcid.org/0000-0002-1939-4332</idno>
              <affiliation ref="#struct-1043327"/>
            </author>
            <editor role="depositor">
              <persName>
                <forename>Kimang</forename>
                <surname>Khun</surname>
              </persName>
              <email type="md5">6805820370ed96ffa0da1e84711e2284</email>
              <email type="domain">gmail.com</email>
            </editor>
          </titleStmt>
          <editionStmt>
            <edition n="v1" type="current">
              <date type="whenSubmitted">2021-06-16 11:43:13</date>
              <date type="whenModified">2025-10-27 10:56:05</date>
              <date type="whenReleased">2021-06-16 15:29:41</date>
              <date type="whenProduced">2021-06-16</date>
              <date type="whenEndEmbargoed">2021-06-16</date>
              <ref type="file" target="https://inria.hal.science/hal-03262006v1/document">
                <date notBefore="2021-06-16"/>
              </ref>
              <ref type="file" subtype="author" n="1" target="https://inria.hal.science/hal-03262006v1/file/gittinsPS.pdf" id="file-3262006-2865834">
                <date notBefore="2021-06-16"/>
              </ref>
              <ref type="externalLink" target="http://arxiv.org/pdf/2106.08771"/>
            </edition>
            <edition n="v2">
              <date type="whenSubmitted">2022-05-02 14:28:01</date>
            </edition>
            <edition n="v3">
              <date type="whenSubmitted">2023-02-09 13:35:40</date>
            </edition>
            <respStmt>
              <resp>contributor</resp>
              <name key="792046">
                <persName>
                  <forename>Kimang</forename>
                  <surname>Khun</surname>
                </persName>
                <email type="md5">6805820370ed96ffa0da1e84711e2284</email>
                <email type="domain">gmail.com</email>
              </name>
            </respStmt>
          </editionStmt>
          <publicationStmt>
            <distributor>CCSD</distributor>
            <idno type="halId">hal-03262006</idno>
            <idno type="halUri">https://inria.hal.science/hal-03262006</idno>
            <idno type="halBibtex">gast:hal-03262006</idno>
            <idno type="halRefHtml">2021</idno>
            <idno type="halRef">2021</idno>
            <availability status="restricted">
              <licence target="https://about.hal.science/hal-authorisation-v1/">HAL Authorization<ref corresp="#file-3262006-2865834"/></licence>
            </availability>
          </publicationStmt>
          <seriesStmt/>
          <notesStmt/>
          <sourceDesc>
            <biblStruct>
              <analytic>
                <title xml:lang="en">Reinforcement Learning for Markovian Bandits: Is Posterior Sampling more Scalable than Optimism?</title>
                <author role="aut">
                  <persName>
                    <forename type="first">Nicolas</forename>
                    <surname>Gast</surname>
                  </persName>
                  <email type="md5">52c7b3fb122eece8a8132fbb2a32bc68</email>
                  <email type="domain">inria.fr</email>
                  <idno type="idhal" notation="string">nicolas-gast</idno>
                  <idno type="idhal" notation="numeric">1247</idno>
                  <idno type="halauthorid" notation="string">17248-1247</idno>
                  <idno type="IDREF">https://www.idref.fr/233247874</idno>
                  <idno type="ORCID">https://orcid.org/0000-0001-6884-8698</idno>
                  <affiliation ref="#struct-1043327"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Bruno</forename>
                    <surname>Gaujal</surname>
                  </persName>
                  <email type="md5">8673cdd3577f13f3e0da99ccfccbabe4</email>
                  <email type="domain">imag.fr</email>
                  <idno type="idhal" notation="string">bruno-gaujal</idno>
                  <idno type="idhal" notation="numeric">11644</idno>
                  <idno type="halauthorid" notation="string">10749-11644</idno>
                  <idno type="ORCID">https://orcid.org/0000-0001-9081-8401</idno>
                  <idno type="IDREF">https://www.idref.fr/074658441</idno>
                  <affiliation ref="#struct-1043327"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Kimang</forename>
                    <surname>Khun</surname>
                  </persName>
                  <email type="md5">6805820370ed96ffa0da1e84711e2284</email>
                  <email type="domain">gmail.com</email>
                  <idno type="idhal" notation="string">kimang-khun</idno>
                  <idno type="idhal" notation="numeric">1102232</idno>
                  <idno type="halauthorid" notation="string">2230586-1102232</idno>
                  <idno type="ORCID">https://orcid.org/0000-0002-1939-4332</idno>
                  <affiliation ref="#struct-1043327"/>
                </author>
              </analytic>
              <monogr>
                <imprint/>
              </monogr>
              <idno type="arxiv">2106.08771</idno>
            </biblStruct>
          </sourceDesc>
          <profileDesc>
            <langUsage>
              <language ident="en">English</language>
            </langUsage>
            <textClass>
              <classCode scheme="halDomain" n="info.info-ai">Computer Science [cs]/Artificial Intelligence [cs.AI]</classCode>
              <classCode scheme="halDomain" n="info.info-lg">Computer Science [cs]/Machine Learning [cs.LG]</classCode>
              <classCode scheme="halTypology" n="UNDEFINED">Preprints, Working Papers, ...</classCode>
              <classCode scheme="halOldTypology" n="UNDEFINED">Preprints, Working Papers, ...</classCode>
              <classCode scheme="halTreeTypology" n="UNDEFINED">Preprints, Working Papers, ...</classCode>
            </textClass>
            <abstract xml:lang="en">
              <p>We study learning algorithms for the classical Markovian bandit problem with discount. We explain how to adapt PSRL [24] and UCRL2 [2] to exploit the problem structure. These variants are called MB-PSRL and MB-UCRL2. While the regret bound and runtime of vanilla implementations of PSRL and UCRL2 are exponential in the number of bandits, we show that the episodic regret of MB-PSRL and MB-UCRL2 isÕ(S √ nK) where K is the number of episodes, n is the number of bandits and S is the number of states of each bandit (the exact bound in S, n and K is given in the paper). Up to a factor √ S, this matches the lower bound of Ω(√ SnK) that we also derive in the paper. MB-PSRL is also computationally efficient: its runtime is linear in the number of bandits. We further show that this linear runtime cannot be achieved by adapting classical non-Bayesian algorithms such as UCRL2 or UCBVI to Markovian bandit problems. Finally, we perform numerical experiments that confirm that MB-PSRL outperforms other existing algorithms in practice, both in terms of regret and of computation time.</p>
            </abstract>
          </profileDesc>
        </biblFull>
      </listBibl>
    </body>
    <back>
      <listOrg type="structures">
        <org type="researchteam" xml:id="struct-1043327" status="VALID">
          <idno type="RNSR">201622036M</idno>
          <idno type="ROR">https://ror.org/056th3f83</idno>
          <orgName>Performance analysis and optimization of LARge Infrastructures and Systems</orgName>
          <orgName type="acronym">POLARIS</orgName>
          <date type="start">2020-01-01</date>
          <date type="end">2026-06-30</date>
          <desc>
            <address>
              <addrLine>Laboratoire LIG - Bâtiment IMAG - 700 avenue Centrale, CS 40700 - 38058 Grenoble cedex 9</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">https://team.inria.fr/polaris/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-2497" type="direct"/>
            <relation active="#struct-300009" type="indirect"/>
            <relation active="#struct-1043301" type="direct"/>
            <relation name="UMR5217" active="#struct-441569" type="indirect"/>
            <relation active="#struct-1042703" type="indirect"/>
            <relation active="#struct-1043329" type="indirect"/>
          </listRelation>
        </org>
        <org type="laboratory" xml:id="struct-2497" status="VALID">
          <idno type="RNSR">199218244V</idno>
          <idno type="ROR">https://ror.org/00n8d6z93</idno>
          <orgName>Centre Inria de l'Université Grenoble Alpes</orgName>
          <desc>
            <address>
              <addrLine>Inovallée655 avenue de l'Europe38330 Montbonnot</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.inria.fr/centre/grenoble</ref>
          </desc>
          <listRelation>
            <relation active="#struct-300009" type="direct"/>
          </listRelation>
        </org>
        <org type="institution" xml:id="struct-300009" status="VALID">
          <idno type="ROR">https://ror.org/02kvxyf05</idno>
          <orgName>Institut National de Recherche en Informatique et en Automatique</orgName>
          <orgName type="acronym">Inria</orgName>
          <desc>
            <address>
              <addrLine>Domaine de VoluceauRocquencourt - BP 10578153 Le Chesnay Cedex</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.inria.fr/en/</ref>
          </desc>
        </org>
        <org type="laboratory" xml:id="struct-1043301" status="VALID">
          <idno type="IdRef">148425550</idno>
          <idno type="ISNI">0000 0001 2286 4035</idno>
          <idno type="RNSR">200711886U</idno>
          <idno type="ROR">https://ror.org/01c8rcg82</idno>
          <orgName>Laboratoire d'Informatique de Grenoble</orgName>
          <orgName type="acronym">LIG</orgName>
          <date type="start">2020-01-01</date>
          <desc>
            <address>
              <addrLine>UMR 5217 - Laboratoire LIG - Bâtiment IMAG - 700 avenue Centrale - Domaine Universitaire de Saint-Martin-d’Hères Adresse postale : CS 40700 - 38058 Grenoble cedex 9Tél. : 04 57 42 14 00</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.liglab.fr/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-300009" type="direct"/>
            <relation name="UMR5217" active="#struct-441569" type="direct"/>
            <relation active="#struct-1042703" type="direct"/>
            <relation active="#struct-1043329" type="direct"/>
          </listRelation>
        </org>
        <org type="regroupinstitution" xml:id="struct-441569" status="VALID">
          <idno type="IdRef">02636817X</idno>
          <idno type="ISNI">0000000122597504</idno>
          <idno type="ROR">https://ror.org/02feahw73</idno>
          <orgName>Centre National de la Recherche Scientifique</orgName>
          <orgName type="acronym">CNRS</orgName>
          <date type="start">1939-10-19</date>
          <desc>
            <address>
              <country key="FR"/>
            </address>
            <ref type="url">https://www.cnrs.fr/</ref>
          </desc>
        </org>
        <org type="regroupinstitution" xml:id="struct-1042703" status="VALID">
          <idno type="IdRef">240648315</idno>
          <idno type="ROR">https://ror.org/02rx3b187</idno>
          <orgName>Université Grenoble Alpes</orgName>
          <orgName type="acronym">UGA</orgName>
          <date type="start">2020-01-01</date>
          <desc>
            <address>
              <addrLine>Adresse CS 40700 - 38058 Grenoble cedex</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.univ-grenoble-alpes.fr</ref>
          </desc>
        </org>
        <org type="institution" xml:id="struct-1043329" status="VALID">
          <idno type="IdRef">026388804</idno>
          <idno type="ROR">https://ror.org/05sbt2524</idno>
          <orgName>Institut polytechnique de Grenoble - Grenoble Institute of Technology</orgName>
          <orgName type="acronym">Grenoble INP</orgName>
          <date type="start">2020-01-01</date>
          <desc>
            <address>
              <addrLine>46 avenue Félix Viallet 38031 Grenoble Cedex 1</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.grenoble-inp.fr/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-1042703" type="direct"/>
          </listRelation>
        </org>
      </listOrg>
    </back>
  </text>
</TEI>