<?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-02951653v4</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="2026-05-19T07:55:28+02:00"/>
      </publicationStmt>
      <sourceDesc>
        <p part="N">HAL API Platform</p>
      </sourceDesc>
    </fileDesc>
  </teiHeader>
  <text>
    <body>
      <listBibl>
        <biblFull>
          <titleStmt>
            <title xml:lang="en">Fundamental Limits of Decentralized Data Shuffling</title>
            <author role="aut">
              <persName>
                <forename type="first">Kai</forename>
                <surname>Wan</surname>
              </persName>
              <idno type="halauthorid">1332666-0</idno>
              <affiliation ref="#struct-86624"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Daniela</forename>
                <surname>Tuninetti</surname>
              </persName>
              <email type="md5">f2f765cc7655cd772975e5f6ad1153da</email>
              <email type="domain">uic.edu</email>
              <idno type="idhal" notation="numeric">947194</idno>
              <idno type="halauthorid" notation="string">741981-947194</idno>
              <affiliation ref="#struct-367491"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Mingyue</forename>
                <surname>Ji</surname>
              </persName>
              <email type="md5">e9e850c72e92ece431f396d0136c5493</email>
              <email type="domain">utah.edu</email>
              <idno type="idhal" notation="numeric">1059997</idno>
              <idno type="halauthorid" notation="string">1332667-1059997</idno>
              <affiliation ref="#struct-300863"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Giuseppe</forename>
                <surname>Caire</surname>
              </persName>
              <idno type="halauthorid">215928-0</idno>
              <affiliation ref="#struct-86624"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Pablo</forename>
                <surname>Piantanida</surname>
              </persName>
              <email type="md5">4d99eb094c1c3424e1f0788dcb8f365d</email>
              <email type="domain">centralesupelec.fr</email>
              <idno type="idhal" notation="string">pablo-piantanida</idno>
              <idno type="idhal" notation="numeric">736967</idno>
              <idno type="halauthorid" notation="string">28498-736967</idno>
              <idno type="ORCID">https://orcid.org/0000-0002-8717-2117</idno>
              <idno type="IDREF">https://www.idref.fr/116306300</idno>
              <affiliation ref="#struct-523655"/>
              <affiliation ref="#struct-411575"/>
            </author>
            <editor role="depositor">
              <persName>
                <forename>Pablo</forename>
                <surname>Piantanida</surname>
              </persName>
              <email type="md5">4d99eb094c1c3424e1f0788dcb8f365d</email>
              <email type="domain">centralesupelec.fr</email>
            </editor>
          </titleStmt>
          <editionStmt>
            <edition n="v1">
              <date type="whenSubmitted">2020-09-28 18:28:26</date>
            </edition>
            <edition n="v2">
              <date type="whenSubmitted">2021-09-21 18:37:23</date>
            </edition>
            <edition n="v3">
              <date type="whenSubmitted">2022-01-19 23:40:20</date>
            </edition>
            <edition n="v4" type="current">
              <date type="whenSubmitted">2023-06-22 03:44:57</date>
              <date type="whenModified">2024-05-24 19:48:04</date>
              <date type="whenReleased">2023-06-22 16:54:14</date>
              <date type="whenProduced">2020</date>
              <date type="whenEndEmbargoed">2023-06-22</date>
              <ref type="file" target="https://centralesupelec.hal.science/hal-02951653v4/document">
                <date notBefore="2023-06-22"/>
              </ref>
              <ref type="file" subtype="author" n="1" target="https://centralesupelec.hal.science/hal-02951653v4/file/Fundamental%20Limits%20of%20Decentralized.pdf" id="file-4137022-3601819">
                <date notBefore="2023-06-22"/>
              </ref>
              <ref type="externalLink" target="http://arxiv.org/pdf/1807.00056"/>
            </edition>
            <respStmt>
              <resp>contributor</resp>
              <name key="329762">
                <persName>
                  <forename>Pablo</forename>
                  <surname>Piantanida</surname>
                </persName>
                <email type="md5">4d99eb094c1c3424e1f0788dcb8f365d</email>
                <email type="domain">centralesupelec.fr</email>
              </name>
            </respStmt>
          </editionStmt>
          <publicationStmt>
            <distributor>CCSD</distributor>
            <idno type="halId">hal-02951653</idno>
            <idno type="halUri">https://centralesupelec.hal.science/hal-02951653</idno>
            <idno type="halBibtex">wan:hal-02951653</idno>
            <idno type="halRefHtml">&lt;i&gt;IEEE Transactions on Information Theory&lt;/i&gt;, 2020, 66 (6), pp.3616-3637. &lt;a target="_blank" href="https://dx.doi.org/10.1109/tit.2020.2966197"&gt;&amp;#x27E8;10.1109/tit.2020.2966197&amp;#x27E9;&lt;/a&gt;</idno>
            <idno type="halRef">IEEE Transactions on Information Theory, 2020, 66 (6), pp.3616-3637. &amp;#x27E8;10.1109/tit.2020.2966197&amp;#x27E9;</idno>
            <availability status="restricted">
              <licence target="https://about.hal.science/hal-authorisation-v1/">HAL Authorization<ref corresp="#file-4137022-3601819"/></licence>
            </availability>
          </publicationStmt>
          <seriesStmt>
            <idno type="stamp" n="CENTRALESUPELEC">Ecole CentraleSupélec</idno>
            <idno type="stamp" n="HUB-IA">Hub Intelligence Artificielle de CentraleSupélec</idno>
          </seriesStmt>
          <notesStmt>
            <note type="audience" n="2">International</note>
            <note type="popular" n="0">No</note>
            <note type="peer" n="1">Yes</note>
          </notesStmt>
          <sourceDesc>
            <biblStruct>
              <analytic>
                <title xml:lang="en">Fundamental Limits of Decentralized Data Shuffling</title>
                <author role="aut">
                  <persName>
                    <forename type="first">Kai</forename>
                    <surname>Wan</surname>
                  </persName>
                  <idno type="halauthorid">1332666-0</idno>
                  <affiliation ref="#struct-86624"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Daniela</forename>
                    <surname>Tuninetti</surname>
                  </persName>
                  <email type="md5">f2f765cc7655cd772975e5f6ad1153da</email>
                  <email type="domain">uic.edu</email>
                  <idno type="idhal" notation="numeric">947194</idno>
                  <idno type="halauthorid" notation="string">741981-947194</idno>
                  <affiliation ref="#struct-367491"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Mingyue</forename>
                    <surname>Ji</surname>
                  </persName>
                  <email type="md5">e9e850c72e92ece431f396d0136c5493</email>
                  <email type="domain">utah.edu</email>
                  <idno type="idhal" notation="numeric">1059997</idno>
                  <idno type="halauthorid" notation="string">1332667-1059997</idno>
                  <affiliation ref="#struct-300863"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Giuseppe</forename>
                    <surname>Caire</surname>
                  </persName>
                  <idno type="halauthorid">215928-0</idno>
                  <affiliation ref="#struct-86624"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Pablo</forename>
                    <surname>Piantanida</surname>
                  </persName>
                  <email type="md5">4d99eb094c1c3424e1f0788dcb8f365d</email>
                  <email type="domain">centralesupelec.fr</email>
                  <idno type="idhal" notation="string">pablo-piantanida</idno>
                  <idno type="idhal" notation="numeric">736967</idno>
                  <idno type="halauthorid" notation="string">28498-736967</idno>
                  <idno type="ORCID">https://orcid.org/0000-0002-8717-2117</idno>
                  <idno type="IDREF">https://www.idref.fr/116306300</idno>
                  <affiliation ref="#struct-523655"/>
                  <affiliation ref="#struct-411575"/>
                </author>
              </analytic>
              <monogr>
                <idno type="halJournalId" status="VALID">5382</idno>
                <idno type="issn">0018-9448</idno>
                <title level="j">IEEE Transactions on Information Theory</title>
                <imprint>
                  <publisher>Institute of Electrical and Electronics Engineers</publisher>
                  <biblScope unit="volume">66</biblScope>
                  <biblScope unit="issue">6</biblScope>
                  <biblScope unit="pp">3616-3637</biblScope>
                  <date type="datePub">2020</date>
                  <date type="dateEpub">2020-01-13</date>
                </imprint>
              </monogr>
              <idno type="arxiv">1807.00056</idno>
              <idno type="doi">10.1109/tit.2020.2966197</idno>
            </biblStruct>
          </sourceDesc>
          <profileDesc>
            <langUsage>
              <language ident="en">English</language>
            </langUsage>
            <textClass>
              <keywords scheme="author">
                <term xml:lang="en">uncoded storage</term>
                <term xml:lang="en">distributed clique covering</term>
                <term xml:lang="en">Decentralized Data shuffling</term>
              </keywords>
              <classCode scheme="halDomain" n="info.info-it">Computer Science [cs]/Information Theory [cs.IT]</classCode>
              <classCode scheme="halTypology" n="ART">Journal articles</classCode>
              <classCode scheme="halOldTypology" n="ART">Journal articles</classCode>
              <classCode scheme="halTreeTypology" n="ART">Journal articles</classCode>
            </textClass>
            <abstract xml:lang="en">
              <p>Data shuffling of training data among different computing nodes (workers) has been identified as a core element to improve the statistical performance of modern large-scale machine learning algorithms. Data shuffling is often considered as one of the most significant bottlenecks in such systems due to the heavy communication load. Under a master-worker architecture (where a master has access to the entire dataset and only communication between the master and the workers is allowed) coding has been recently proved to considerably reduce the communication load. This work considers a different communication paradigm referred to as decentralized data shuffling, where workers are allowed to communicate with one another via a shared link. The decentralized data shuffling problem has two phases: workers communicate with each other during the data shuffling phase, and then workers update their stored content during the storage phase. The main challenge is to derive novel converse bounds and achievable schemes for decentralized data shuffling by considering the asymmetry of the workers' storages (i.e., workers are constrained to store different files in their storages based on the problem setting), in order to characterize the fundamental limits of this problem. For the case of uncoded storage (i.e., each worker directly stores a subset of bits of the dataset), this paper proposes converse and achievable bounds (based on distributed interference alignment and distributed clique-covering strategies) that are within a factor of 3/2 of one another. The proposed schemes are also exactly optimal under the constraint of uncoded storage for either large storage size or at most four workers in the system.</p>
            </abstract>
          </profileDesc>
        </biblFull>
      </listBibl>
    </body>
    <back>
      <listOrg type="structures">
        <org type="institution" xml:id="struct-86624" status="VALID">
          <idno type="ROR">https://ror.org/03v4gjf40</idno>
          <orgName>Technical University of Berlin / Technische Universität Berlin</orgName>
          <orgName type="acronym">TUB</orgName>
          <date type="start">1946-01-01</date>
          <desc>
            <address>
              <addrLine>Straße des 17. Juni 135 10623 Berlin</addrLine>
              <country key="DE"/>
            </address>
            <ref type="url">https://www.tu.berlin/</ref>
          </desc>
        </org>
        <org type="institution" xml:id="struct-367491" status="VALID">
          <idno type="ROR">https://ror.org/02mpq6x41</idno>
          <orgName>University of Illinois [Chicago]</orgName>
          <orgName type="acronym">UIC</orgName>
          <desc>
            <address>
              <addrLine>1200 West Harrison St., Chicago, Illinois 60607</addrLine>
              <country key="US"/>
            </address>
            <ref type="url">http://www.uic.edu</ref>
          </desc>
          <listRelation>
            <relation active="#struct-325015" type="direct"/>
          </listRelation>
        </org>
        <org type="institution" xml:id="struct-300863" status="VALID">
          <idno type="ROR">https://ror.org/03r0ha626</idno>
          <orgName>University of Utah</orgName>
          <desc>
            <address>
              <addrLine>201 Presidents Cir, Salt Lake City, UT 84112</addrLine>
              <country key="US"/>
            </address>
            <ref type="url">http://www.utah.edu/</ref>
          </desc>
        </org>
        <org type="laboratory" xml:id="struct-523655" status="VALID">
          <orgName>Montreal Institute for Learning Algorithms  [Montréal]</orgName>
          <orgName type="acronym">MILA</orgName>
          <desc>
            <address>
              <addrLine>6666 St-Urbain, #200, Montreal, QC, H2S 3H1, Canada</addrLine>
              <country key="CA"/>
            </address>
            <ref type="url">https://mila.quebec/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-231516" type="direct"/>
            <relation active="#struct-302452" type="indirect"/>
          </listRelation>
        </org>
        <org type="institution" xml:id="struct-411575" status="VALID">
          <idno type="IdRef">184443237</idno>
          <idno type="ROR">https://ror.org/019tcpt25</idno>
          <orgName>CentraleSupélec</orgName>
          <desc>
            <address>
              <addrLine>3, rue Joliot Curie,Plateau de Moulon,91192 GIF-SUR-YVETTE Cedex</addrLine>
              <country key="FR"/>
            </address>
            <ref type="url">http://www.centralesupelec.fr</ref>
          </desc>
        </org>
        <org type="regroupinstitution" xml:id="struct-325015" status="VALID">
          <idno type="ROR">https://ror.org/05e94g991</idno>
          <orgName>University of Illinois System</orgName>
          <desc>
            <address>
              <addrLine>108 Henry Administration Bldg.Urbana, IL 61801</addrLine>
              <country key="US"/>
            </address>
            <ref type="url">https://www.uillinois.edu/</ref>
          </desc>
        </org>
        <org type="institution" xml:id="struct-231516" status="VALID">
          <orgName>Centre de Recherches Mathématiques [Montréal]</orgName>
          <orgName type="acronym">CRM</orgName>
          <date type="start">1968-01-01</date>
          <desc>
            <address>
              <addrLine>Université de Montréal, Pavillon André-Aisenstadt, 2920, Chemin de la tour, bur. 5357 Montréal (Québec) H3T 1J4</addrLine>
              <country key="CA"/>
            </address>
            <ref type="url">http://www.crm.umontreal.ca/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-302452" type="direct"/>
          </listRelation>
        </org>
        <org type="regroupinstitution" xml:id="struct-302452" status="VALID">
          <idno type="ROR">https://ror.org/0161xgx34</idno>
          <orgName>Université de Montréal</orgName>
          <orgName type="acronym">UdeM</orgName>
          <desc>
            <address>
              <addrLine>2900 Boulevard Edouard-Montpetit, Montréal, QC H3T 1J4</addrLine>
              <country key="CA"/>
            </address>
            <ref type="url">https://www.umontreal.ca/</ref>
          </desc>
        </org>
      </listOrg>
    </back>
  </text>
</TEI>