<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="sen-etal-2021-university">
    <titleInfo>
        <title>The University of Edinburgh’s Submission to the IWSLT21 Simultaneous Translation Task</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Sukanta</namePart>
        <namePart type="family">Sen</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ulrich</namePart>
        <namePart type="family">Germann</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Barry</namePart>
        <namePart type="family">Haddow</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2021-08</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 18th International Conference on Spoken Language Translation (IWSLT 2021)</title>
        </titleInfo>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Bangkok, Thailand (online)</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
    </relatedItem>
    <abstract>We describe our submission to the IWSLT 2021 shared task on simultaneous text-to-text English-German translation. Our system is based on the re-translation approach where the agent re-translates the whole source prefix each time it receives a new source token. This approach has the advantage of being able to use a standard neural machine translation (NMT) inference engine with beam search, however, there is a risk that incompatibility between successive re-translations will degrade the output. To improve the quality of the translations, we experiment with various approaches: we use a fixed size wait at the beginning of the sentence, we use a language model score to detect translatable units, and we apply dynamic masking to determine when the translation is unstable. We find that a combination of dynamic masking and language model score obtains the best latency-quality trade-off.</abstract>
    <identifier type="citekey">sen-etal-2021-university</identifier>
    <identifier type="doi">10.18653/v1/2021.iwslt-1.4</identifier>
    <location>
        <url>https://aclanthology.org/2021.iwslt-1.4</url>
    </location>
    <part>
        <date>2021-08</date>
        <extent unit="page">
            <start>46</start>
            <end>51</end>
        </extent>
    </part>
</mods>
</modsCollection>

%0 Conference Proceedings
%T The University of Edinburgh’s Submission to the IWSLT21 Simultaneous Translation Task
%A Sen, Sukanta
%A Germann, Ulrich
%A Haddow, Barry
%S Proceedings of the 18th International Conference on Spoken Language Translation (IWSLT 2021)
%D 2021
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand (online)
%F sen-etal-2021-university
%X We describe our submission to the IWSLT 2021 shared task on simultaneous text-to-text English-German translation. Our system is based on the re-translation approach where the agent re-translates the whole source prefix each time it receives a new source token. This approach has the advantage of being able to use a standard neural machine translation (NMT) inference engine with beam search, however, there is a risk that incompatibility between successive re-translations will degrade the output. To improve the quality of the translations, we experiment with various approaches: we use a fixed size wait at the beginning of the sentence, we use a language model score to detect translatable units, and we apply dynamic masking to determine when the translation is unstable. We find that a combination of dynamic masking and language model score obtains the best latency-quality trade-off.
%R 10.18653/v1/2021.iwslt-1.4
%U https://aclanthology.org/2021.iwslt-1.4
%U https://doi.org/10.18653/v1/2021.iwslt-1.4
%P 46-51

Markdown (Informal)

[The University of Edinburgh’s Submission to the IWSLT21 Simultaneous Translation Task](https://aclanthology.org/2021.iwslt-1.4) (Sen et al., IWSLT 2021)

• The University of Edinburgh’s Submission to the IWSLT21 Simultaneous Translation Task (Sen et al., IWSLT 2021)

ACL

• Sukanta Sen, Ulrich Germann, and Barry Haddow. 2021. The University of Edinburgh’s Submission to the IWSLT21 Simultaneous Translation Task. In Proceedings of the 18th International Conference on Spoken Language Translation (IWSLT 2021), pages 46–51, Bangkok, Thailand (online). Association for Computational Linguistics.