@inproceedings{chang-etal-2022-anticipation,
title = "Anticipation-Free Training for Simultaneous Machine Translation",
author = "Chang, Chih-Chiang and
Chuang, Shun-Po and
Lee, Hung-yi",
editor = "Salesky, Elizabeth and
Federico, Marcello and
Costa-juss{\`a}, Marta",
booktitle = "Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022)",
month = may,
year = "2022",
address = "Dublin, Ireland (in-person and online)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.iwslt-1.5/",
doi = "10.18653/v1/2022.iwslt-1.5",
pages = "43--61",
abstract = "Simultaneous machine translation (SimulMT) speeds up the translation process by starting to translate before the source sentence is completely available. It is difficult due to limited context and word order difference between languages. Existing methods increase latency or introduce adaptive read-write policies for SimulMT models to handle local reordering and improve translation quality. However, the long-distance reordering would make the SimulMT models learn translation mistakenly. Specifically, the model may be forced to predict target tokens when the corresponding source tokens have not been read. This leads to aggressive anticipation during inference, resulting in the hallucination phenomenon. To mitigate this problem, we propose a new framework that decompose the translation process into the monotonic translation step and the reordering step, and we model the latter by the auxiliary sorting network (ASN). The ASN rearranges the hidden states to match the order in the target language, so that the SimulMT model could learn to translate more reasonably. The entire model is optimized end-to-end and does not rely on external aligners or data. During inference, ASN is removed to achieve streaming. Experiments show the proposed framework could outperform previous methods with less latency."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chang-etal-2022-anticipation">
<titleInfo>
<title>Anticipation-Free Training for Simultaneous Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chih-Chiang</namePart>
<namePart type="family">Chang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shun-Po</namePart>
<namePart type="family">Chuang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hung-yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2022-05</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Elizabeth</namePart>
<namePart type="family">Salesky</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marcello</namePart>
<namePart type="family">Federico</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Marta</namePart>
<namePart type="family">Costa-jussà</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Dublin, Ireland (in-person and online)</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Simultaneous machine translation (SimulMT) speeds up the translation process by starting to translate before the source sentence is completely available. It is difficult due to limited context and word order difference between languages. Existing methods increase latency or introduce adaptive read-write policies for SimulMT models to handle local reordering and improve translation quality. However, the long-distance reordering would make the SimulMT models learn translation mistakenly. Specifically, the model may be forced to predict target tokens when the corresponding source tokens have not been read. This leads to aggressive anticipation during inference, resulting in the hallucination phenomenon. To mitigate this problem, we propose a new framework that decompose the translation process into the monotonic translation step and the reordering step, and we model the latter by the auxiliary sorting network (ASN). The ASN rearranges the hidden states to match the order in the target language, so that the SimulMT model could learn to translate more reasonably. The entire model is optimized end-to-end and does not rely on external aligners or data. During inference, ASN is removed to achieve streaming. Experiments show the proposed framework could outperform previous methods with less latency.</abstract>
<identifier type="citekey">chang-etal-2022-anticipation</identifier>
<identifier type="doi">10.18653/v1/2022.iwslt-1.5</identifier>
<location>
<url>https://aclanthology.org/2022.iwslt-1.5/</url>
</location>
<part>
<date>2022-05</date>
<extent unit="page">
<start>43</start>
<end>61</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Anticipation-Free Training for Simultaneous Machine Translation
%A Chang, Chih-Chiang
%A Chuang, Shun-Po
%A Lee, Hung-yi
%Y Salesky, Elizabeth
%Y Federico, Marcello
%Y Costa-jussà, Marta
%S Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022)
%D 2022
%8 May
%I Association for Computational Linguistics
%C Dublin, Ireland (in-person and online)
%F chang-etal-2022-anticipation
%X Simultaneous machine translation (SimulMT) speeds up the translation process by starting to translate before the source sentence is completely available. It is difficult due to limited context and word order difference between languages. Existing methods increase latency or introduce adaptive read-write policies for SimulMT models to handle local reordering and improve translation quality. However, the long-distance reordering would make the SimulMT models learn translation mistakenly. Specifically, the model may be forced to predict target tokens when the corresponding source tokens have not been read. This leads to aggressive anticipation during inference, resulting in the hallucination phenomenon. To mitigate this problem, we propose a new framework that decompose the translation process into the monotonic translation step and the reordering step, and we model the latter by the auxiliary sorting network (ASN). The ASN rearranges the hidden states to match the order in the target language, so that the SimulMT model could learn to translate more reasonably. The entire model is optimized end-to-end and does not rely on external aligners or data. During inference, ASN is removed to achieve streaming. Experiments show the proposed framework could outperform previous methods with less latency.
%R 10.18653/v1/2022.iwslt-1.5
%U https://aclanthology.org/2022.iwslt-1.5/
%U https://doi.org/10.18653/v1/2022.iwslt-1.5
%P 43-61
Markdown (Informal)
[Anticipation-Free Training for Simultaneous Machine Translation](https://aclanthology.org/2022.iwslt-1.5/) (Chang et al., IWSLT 2022)
ACL
- Chih-Chiang Chang, Shun-Po Chuang, and Hung-yi Lee. 2022. Anticipation-Free Training for Simultaneous Machine Translation. In Proceedings of the 19th International Conference on Spoken Language Translation (IWSLT 2022), pages 43–61, Dublin, Ireland (in-person and online). Association for Computational Linguistics.