@inproceedings{miao-etal-2021-prevent,
title = "Prevent the Language Model from being Overconfident in Neural Machine Translation",
author = "Miao, Mengqi and
Meng, Fandong and
Liu, Yijin and
Zhou, Xiao-Hua and
Zhou, Jie",
editor = "Zong, Chengqing and
Xia, Fei and
Li, Wenjie and
Navigli, Roberto",
booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)",
month = aug,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.acl-long.268",
doi = "10.18653/v1/2021.acl-long.268",
pages = "3456--3468",
abstract = "The Neural Machine Translation (NMT) model is essentially a joint language model conditioned on both the source sentence and partial translation. Therefore, the NMT model naturally involves the mechanism of the Language Model (LM) that predicts the next token only based on partial translation. Despite its success, NMT still suffers from the hallucination problem, generating fluent but inadequate translations. The main reason is that NMT pays excessive attention to the partial translation while neglecting the source sentence to some extent, namely overconfidence of the LM. Accordingly, we define the Margin between the NMT and the LM, calculated by subtracting the predicted probability of the LM from that of the NMT model for each token. The Margin is negatively correlated to the overconfidence degree of the LM. Based on the property, we propose a Margin-based Token-level Objective (MTO) and a Margin-based Sentence-level Objective (MSO) to maximize the Margin for preventing the LM from being overconfident. Experiments on WMT14 English-to-German, WMT19 Chinese-to-English, and WMT14 English-to-French translation tasks demonstrate the effectiveness of our approach, with 1.36, 1.50, and 0.63 BLEU improvements, respectively, compared to the Transformer baseline. The human evaluation further verifies that our approaches improve translation adequacy as well as fluency.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="miao-etal-2021-prevent">
<titleInfo>
<title>Prevent the Language Model from being Overconfident in Neural Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mengqi</namePart>
<namePart type="family">Miao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fandong</namePart>
<namePart type="family">Meng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yijin</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiao-Hua</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jie</namePart>
<namePart type="family">Zhou</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 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chengqing</namePart>
<namePart type="family">Zong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fei</namePart>
<namePart type="family">Xia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wenjie</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Roberto</namePart>
<namePart type="family">Navigli</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The Neural Machine Translation (NMT) model is essentially a joint language model conditioned on both the source sentence and partial translation. Therefore, the NMT model naturally involves the mechanism of the Language Model (LM) that predicts the next token only based on partial translation. Despite its success, NMT still suffers from the hallucination problem, generating fluent but inadequate translations. The main reason is that NMT pays excessive attention to the partial translation while neglecting the source sentence to some extent, namely overconfidence of the LM. Accordingly, we define the Margin between the NMT and the LM, calculated by subtracting the predicted probability of the LM from that of the NMT model for each token. The Margin is negatively correlated to the overconfidence degree of the LM. Based on the property, we propose a Margin-based Token-level Objective (MTO) and a Margin-based Sentence-level Objective (MSO) to maximize the Margin for preventing the LM from being overconfident. Experiments on WMT14 English-to-German, WMT19 Chinese-to-English, and WMT14 English-to-French translation tasks demonstrate the effectiveness of our approach, with 1.36, 1.50, and 0.63 BLEU improvements, respectively, compared to the Transformer baseline. The human evaluation further verifies that our approaches improve translation adequacy as well as fluency.</abstract>
<identifier type="citekey">miao-etal-2021-prevent</identifier>
<identifier type="doi">10.18653/v1/2021.acl-long.268</identifier>
<location>
<url>https://aclanthology.org/2021.acl-long.268</url>
</location>
<part>
<date>2021-08</date>
<extent unit="page">
<start>3456</start>
<end>3468</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Prevent the Language Model from being Overconfident in Neural Machine Translation
%A Miao, Mengqi
%A Meng, Fandong
%A Liu, Yijin
%A Zhou, Xiao-Hua
%A Zhou, Jie
%Y Zong, Chengqing
%Y Xia, Fei
%Y Li, Wenjie
%Y Navigli, Roberto
%S Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
%D 2021
%8 August
%I Association for Computational Linguistics
%C Online
%F miao-etal-2021-prevent
%X The Neural Machine Translation (NMT) model is essentially a joint language model conditioned on both the source sentence and partial translation. Therefore, the NMT model naturally involves the mechanism of the Language Model (LM) that predicts the next token only based on partial translation. Despite its success, NMT still suffers from the hallucination problem, generating fluent but inadequate translations. The main reason is that NMT pays excessive attention to the partial translation while neglecting the source sentence to some extent, namely overconfidence of the LM. Accordingly, we define the Margin between the NMT and the LM, calculated by subtracting the predicted probability of the LM from that of the NMT model for each token. The Margin is negatively correlated to the overconfidence degree of the LM. Based on the property, we propose a Margin-based Token-level Objective (MTO) and a Margin-based Sentence-level Objective (MSO) to maximize the Margin for preventing the LM from being overconfident. Experiments on WMT14 English-to-German, WMT19 Chinese-to-English, and WMT14 English-to-French translation tasks demonstrate the effectiveness of our approach, with 1.36, 1.50, and 0.63 BLEU improvements, respectively, compared to the Transformer baseline. The human evaluation further verifies that our approaches improve translation adequacy as well as fluency.
%R 10.18653/v1/2021.acl-long.268
%U https://aclanthology.org/2021.acl-long.268
%U https://doi.org/10.18653/v1/2021.acl-long.268
%P 3456-3468
Markdown (Informal)
[Prevent the Language Model from being Overconfident in Neural Machine Translation](https://aclanthology.org/2021.acl-long.268) (Miao et al., ACL-IJCNLP 2021)
ACL
- Mengqi Miao, Fandong Meng, Yijin Liu, Xiao-Hua Zhou, and Jie Zhou. 2021. Prevent the Language Model from being Overconfident in Neural Machine Translation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 3456–3468, Online. Association for Computational Linguistics.