@inproceedings{he-etal-2022-bridging,
title = "Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation",
author = "He, Zhiwei and
Wang, Xing and
Wang, Rui and
Shi, Shuming and
Tu, Zhaopeng",
editor = "Muresan, Smaranda and
Nakov, Preslav and
Villavicencio, Aline",
booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = may,
year = "2022",
address = "Dublin, Ireland",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.acl-long.456",
doi = "10.18653/v1/2022.acl-long.456",
pages = "6611--6623",
abstract = "Back-translation is a critical component of Unsupervised Neural Machine Translation (UNMT), which generates pseudo parallel data from target monolingual data. A UNMT model is trained on the pseudo parallel data with $\text{\bf translated source}$, and translates $\text{\bf natural source}$ sentences in inference. The source discrepancy between training and inference hinders the translation performance of UNMT models. By carefully designing experiments, we identify two representative characteristics of the data gap in source: (1) $\text{\textit{style gap}}$ (i.e., translated vs. natural text style) that leads to poor generalization capability; (2) $\text{\textit{content gap}}$ that induces the model to produce hallucination content biased towards the target language. To narrow the data gap, we propose an online self-training approach, which simultaneously uses the pseudo parallel data $\{$natural source, translated target$\}$ to mimic the inference scenario. Experimental results on several widely-used language pairs show that our approach outperforms two strong baselines (XLM and MASS) by remedying the style and content gaps.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="he-etal-2022-bridging">
<titleInfo>
<title>Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhiwei</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xing</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rui</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shuming</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhaopeng</namePart>
<namePart type="family">Tu</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 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Smaranda</namePart>
<namePart type="family">Muresan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Preslav</namePart>
<namePart type="family">Nakov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Aline</namePart>
<namePart type="family">Villavicencio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Dublin, Ireland</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Back-translation is a critical component of Unsupervised Neural Machine Translation (UNMT), which generates pseudo parallel data from target monolingual data. A UNMT model is trained on the pseudo parallel data with \texttranslated source, and translates \textnatural source sentences in inference. The source discrepancy between training and inference hinders the translation performance of UNMT models. By carefully designing experiments, we identify two representative characteristics of the data gap in source: (1) \textstyle gap (i.e., translated vs. natural text style) that leads to poor generalization capability; (2) \textcontent gap that induces the model to produce hallucination content biased towards the target language. To narrow the data gap, we propose an online self-training approach, which simultaneously uses the pseudo parallel data {natural source, translated target} to mimic the inference scenario. Experimental results on several widely-used language pairs show that our approach outperforms two strong baselines (XLM and MASS) by remedying the style and content gaps.</abstract>
<identifier type="citekey">he-etal-2022-bridging</identifier>
<identifier type="doi">10.18653/v1/2022.acl-long.456</identifier>
<location>
<url>https://aclanthology.org/2022.acl-long.456</url>
</location>
<part>
<date>2022-05</date>
<extent unit="page">
<start>6611</start>
<end>6623</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation
%A He, Zhiwei
%A Wang, Xing
%A Wang, Rui
%A Shi, Shuming
%A Tu, Zhaopeng
%Y Muresan, Smaranda
%Y Nakov, Preslav
%Y Villavicencio, Aline
%S Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2022
%8 May
%I Association for Computational Linguistics
%C Dublin, Ireland
%F he-etal-2022-bridging
%X Back-translation is a critical component of Unsupervised Neural Machine Translation (UNMT), which generates pseudo parallel data from target monolingual data. A UNMT model is trained on the pseudo parallel data with \texttranslated source, and translates \textnatural source sentences in inference. The source discrepancy between training and inference hinders the translation performance of UNMT models. By carefully designing experiments, we identify two representative characteristics of the data gap in source: (1) \textstyle gap (i.e., translated vs. natural text style) that leads to poor generalization capability; (2) \textcontent gap that induces the model to produce hallucination content biased towards the target language. To narrow the data gap, we propose an online self-training approach, which simultaneously uses the pseudo parallel data {natural source, translated target} to mimic the inference scenario. Experimental results on several widely-used language pairs show that our approach outperforms two strong baselines (XLM and MASS) by remedying the style and content gaps.
%R 10.18653/v1/2022.acl-long.456
%U https://aclanthology.org/2022.acl-long.456
%U https://doi.org/10.18653/v1/2022.acl-long.456
%P 6611-6623
Markdown (Informal)
[Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation](https://aclanthology.org/2022.acl-long.456) (He et al., ACL 2022)
ACL