<?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-02135081</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="2026-05-03T20:19:02+02:00"/>
      </publicationStmt>
      <sourceDesc>
        <p part="N">HAL API Platform</p>
      </sourceDesc>
    </fileDesc>
  </teiHeader>
  <text>
    <body>
      <listBibl>
        <biblFull>
          <titleStmt>
            <title xml:lang="en">Privacy-preserving Neural Representations of Text</title>
            <author role="aut">
              <persName>
                <forename type="first">Maximin</forename>
                <surname>Coavoux</surname>
              </persName>
              <email type="md5">1f6ec3a214411c079c0b414954342746</email>
              <email type="domain">gmail.com</email>
              <idno type="idhal" notation="string">maximin-coavoux</idno>
              <idno type="idhal" notation="numeric">13643</idno>
              <idno type="halauthorid" notation="string">50214-13643</idno>
              <idno type="ORCID">https://orcid.org/0000-0003-4089-4558</idno>
              <affiliation ref="#struct-182763"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Shashi</forename>
                <surname>Narayan</surname>
              </persName>
              <idno type="idhal" notation="numeric">767539</idno>
              <idno type="halauthorid" notation="string">681037-767539</idno>
              <idno type="IDREF">https://www.idref.fr/182508757</idno>
              <affiliation ref="#struct-182763"/>
            </author>
            <author role="aut">
              <persName>
                <forename type="first">Shay B.</forename>
                <surname>Cohen</surname>
              </persName>
              <idno type="halauthorid">1588182-0</idno>
              <affiliation ref="#struct-182763"/>
            </author>
            <editor role="depositor">
              <persName>
                <forename>Maximin</forename>
                <surname>Coavoux</surname>
              </persName>
              <email type="md5">1f6ec3a214411c079c0b414954342746</email>
              <email type="domain">gmail.com</email>
            </editor>
          </titleStmt>
          <editionStmt>
            <edition n="v1" type="current">
              <date type="whenSubmitted">2019-06-07 00:16:02</date>
              <date type="whenModified">2025-02-19 15:45:31</date>
              <date type="whenReleased">2019-06-11 12:43:47</date>
              <date type="whenProduced">2018-11-02</date>
              <date type="whenEndEmbargoed">2019-06-07</date>
              <ref type="file" target="https://hal.science/hal-02135081v1/document">
                <date notBefore="2019-06-07"/>
              </ref>
              <ref type="file" subtype="greenPublisher" n="1" target="https://hal.science/hal-02135081v1/file/D18-1001.pdf" id="file-2150074-2144119">
                <date notBefore="2019-06-07"/>
              </ref>
            </edition>
            <respStmt>
              <resp>contributor</resp>
              <name key="335200">
                <persName>
                  <forename>Maximin</forename>
                  <surname>Coavoux</surname>
                </persName>
                <email type="md5">1f6ec3a214411c079c0b414954342746</email>
                <email type="domain">gmail.com</email>
              </name>
            </respStmt>
          </editionStmt>
          <publicationStmt>
            <distributor>CCSD</distributor>
            <idno type="halId">hal-02135081</idno>
            <idno type="halUri">https://hal.science/hal-02135081</idno>
            <idno type="halBibtex">coavoux:hal-02135081</idno>
            <idno type="halRefHtml">&lt;i&gt;2018 Conference on Empirical Methods in Natural Language Processing&lt;/i&gt;, Nov 2018, Brussels, Belgium. pp.1--10</idno>
            <idno type="halRef">2018 Conference on Empirical Methods in Natural Language Processing, Nov 2018, Brussels, Belgium. pp.1--10</idno>
            <availability status="restricted">
              <licence target="https://about.hal.science/hal-authorisation-v1/">HAL Authorization<ref corresp="#file-2150074-2144119"/></licence>
            </availability>
          </publicationStmt>
          <seriesStmt/>
          <notesStmt>
            <note type="audience" n="2">International</note>
            <note type="invited" n="0">No</note>
            <note type="popular" n="0">No</note>
            <note type="peer" n="1">Yes</note>
            <note type="proceedings" n="0">No</note>
          </notesStmt>
          <sourceDesc>
            <biblStruct>
              <analytic>
                <title xml:lang="en">Privacy-preserving Neural Representations of Text</title>
                <author role="aut">
                  <persName>
                    <forename type="first">Maximin</forename>
                    <surname>Coavoux</surname>
                  </persName>
                  <email type="md5">1f6ec3a214411c079c0b414954342746</email>
                  <email type="domain">gmail.com</email>
                  <idno type="idhal" notation="string">maximin-coavoux</idno>
                  <idno type="idhal" notation="numeric">13643</idno>
                  <idno type="halauthorid" notation="string">50214-13643</idno>
                  <idno type="ORCID">https://orcid.org/0000-0003-4089-4558</idno>
                  <affiliation ref="#struct-182763"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Shashi</forename>
                    <surname>Narayan</surname>
                  </persName>
                  <idno type="idhal" notation="numeric">767539</idno>
                  <idno type="halauthorid" notation="string">681037-767539</idno>
                  <idno type="IDREF">https://www.idref.fr/182508757</idno>
                  <affiliation ref="#struct-182763"/>
                </author>
                <author role="aut">
                  <persName>
                    <forename type="first">Shay B.</forename>
                    <surname>Cohen</surname>
                  </persName>
                  <idno type="halauthorid">1588182-0</idno>
                  <affiliation ref="#struct-182763"/>
                </author>
              </analytic>
              <monogr>
                <title level="m">Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing</title>
                <meeting>
                  <title>2018 Conference on Empirical Methods in Natural Language Processing</title>
                  <date type="start">2018-11-02</date>
                  <settlement>Brussels</settlement>
                  <country key="BE">Belgium</country>
                </meeting>
                <imprint>
                  <publisher>Association for Computational Linguistics</publisher>
                  <biblScope unit="pp">1--10</biblScope>
                  <date type="datePub">2018</date>
                </imprint>
              </monogr>
              <ref target="https://www.aclweb.org/anthology/D18-1001" type="seeAlso"/>
            </biblStruct>
          </sourceDesc>
          <profileDesc>
            <langUsage>
              <language ident="en">English</language>
            </langUsage>
            <textClass>
              <classCode scheme="halDomain" n="info.info-cl">Computer Science [cs]/Computation and Language [cs.CL]</classCode>
              <classCode scheme="halTypology" n="COMM">Conference papers</classCode>
              <classCode scheme="halOldTypology" n="COMM">Conference papers</classCode>
              <classCode scheme="halTreeTypology" n="COMM">Conference papers</classCode>
            </textClass>
            <abstract xml:lang="en">
              <p>This article deals with adversarial attacks towards deep learning systems for Natural Language Processing (NLP), in the context of privacy protection. We study a specific type of attack: an attacker eavesdrops on the hidden representations of a neural text classifier and tries to recover information about the input text. Such scenario may arise in situations when the computation of a neural network is shared across multiple devices, e.g. some hidden representation is computed by a user's device and sent to a cloud-based model. We measure the privacy of a hidden representation by the ability of an attacker to predict accurately specific private information from it and characterize the tradeoff between the privacy and the utility of neural representations. Finally, we propose several defense methods based on modified training objectives and show that they improve the privacy of neural representations.</p>
            </abstract>
          </profileDesc>
        </biblFull>
      </listBibl>
    </body>
    <back>
      <listOrg type="structures">
        <org type="laboratory" xml:id="struct-182763" status="VALID">
          <orgName>School of Informatics [Edimbourg]</orgName>
          <desc>
            <address>
              <addrLine>Informatics Forum 10 Crichton Street Edinburgh EH8 9AB</addrLine>
              <country key="GB"/>
            </address>
            <ref type="url">http://www.ed.ac.uk/schools-departments/informatics/</ref>
          </desc>
          <listRelation>
            <relation active="#struct-61277" type="direct"/>
          </listRelation>
        </org>
        <org type="institution" xml:id="struct-61277" status="VALID">
          <idno type="IdRef">02663483X</idno>
          <idno type="ROR">https://ror.org/01nrxwf90</idno>
          <orgName>The University of Edinburgh</orgName>
          <date type="start">1583-01-01</date>
          <desc>
            <address>
              <addrLine>Old College South Bridge Edinburgh EH8 9YL</addrLine>
              <country key="GB"/>
            </address>
            <ref type="url">http://www.ed.ac.uk/home</ref>
          </desc>
        </org>
      </listOrg>
    </back>
  </text>
</TEI>