Conference paper Open Access

Domain Adaptation of Document-Level NMT in IWSLT19

Popel, Martin; Federmann, Christian


MARC21 XML Export

<?xml version='1.0' encoding='UTF-8'?>
<record xmlns="http://www.loc.gov/MARC21/slim">
  <leader>00000nam##2200000uu#4500</leader>
  <datafield tag="041" ind1=" " ind2=" ">
    <subfield code="a">eng</subfield>
  </datafield>
  <controlfield tag="005">20200120174307.0</controlfield>
  <controlfield tag="001">3525546</controlfield>
  <datafield tag="700" ind1=" " ind2=" ">
    <subfield code="u">Microsoft, 1 Microsoft Way, Redmond, WA 98121, USA</subfield>
    <subfield code="a">Federmann, Christian</subfield>
  </datafield>
  <datafield tag="856" ind1="4" ind2=" ">
    <subfield code="s">153535</subfield>
    <subfield code="z">md5:7d9efc618d2862f227e4c86d309c2590</subfield>
    <subfield code="u">https://zenodo.org/record/3525546/files/IWSLT2019_paper_35.pdf</subfield>
  </datafield>
  <datafield tag="542" ind1=" " ind2=" ">
    <subfield code="l">open</subfield>
  </datafield>
  <datafield tag="260" ind1=" " ind2=" ">
    <subfield code="c">2019-11-02</subfield>
  </datafield>
  <datafield tag="909" ind1="C" ind2="O">
    <subfield code="p">openaire</subfield>
    <subfield code="p">user-iwslt2019</subfield>
    <subfield code="o">oai:zenodo.org:3525546</subfield>
  </datafield>
  <datafield tag="100" ind1=" " ind2=" ">
    <subfield code="u">Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics, Malostranské námeˇstí 25, 118 00 Prague, Czech Republic &amp; Microsoft, 1 Microsoft Way, Redmond, WA 98121, USA</subfield>
    <subfield code="a">Popel, Martin</subfield>
  </datafield>
  <datafield tag="245" ind1=" " ind2=" ">
    <subfield code="a">Domain Adaptation of Document-Level NMT in IWSLT19</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">user-iwslt2019</subfield>
  </datafield>
  <datafield tag="540" ind1=" " ind2=" ">
    <subfield code="u">https://creativecommons.org/licenses/by/4.0/legalcode</subfield>
    <subfield code="a">Creative Commons Attribution 4.0 International</subfield>
  </datafield>
  <datafield tag="650" ind1="1" ind2="7">
    <subfield code="a">cc-by</subfield>
    <subfield code="2">opendefinition.org</subfield>
  </datafield>
  <datafield tag="520" ind1=" " ind2=" ">
    <subfield code="a">&lt;p&gt;We describe our four NMT systems submitted to the IWSLT19 shared task in English&amp;rarr;Czech text-to-text translation of TED talks. The goal of this study is to understand the interactions between document-level NMT and domain adaptation. All our systems are based on the Transformer model implemented in the Tensor2Tensor framework. Two of the systems serve as baselines, which are not adapted to the TED talks domain: SENTBASE&amp;nbsp;is trained on single sen- tences, DOCBASE&amp;nbsp;on multi-sentence (document-level) sequences. The other two submitted systems are adapted to TED talks: SENTFINE&amp;nbsp;is fine-tuned on single sentences, DOCFINE&amp;nbsp;is fine-tuned on multi-sentence sequences. We present both automatic-metrics evaluation and manual analysis of the translation quality, focusing on the differences between the four systems.&lt;/p&gt;</subfield>
  </datafield>
  <datafield tag="773" ind1=" " ind2=" ">
    <subfield code="n">doi</subfield>
    <subfield code="i">isVersionOf</subfield>
    <subfield code="a">10.5281/zenodo.3525545</subfield>
  </datafield>
  <datafield tag="024" ind1=" " ind2=" ">
    <subfield code="a">10.5281/zenodo.3525546</subfield>
    <subfield code="2">doi</subfield>
  </datafield>
  <datafield tag="980" ind1=" " ind2=" ">
    <subfield code="a">publication</subfield>
    <subfield code="b">conferencepaper</subfield>
  </datafield>
</record>
81
96
views
downloads
All versions This version
Views 8181
Downloads 9696
Data volume 14.7 MB14.7 MB
Unique views 7474
Unique downloads 9292

Share

Cite as