@inproceedings{lu-etal-2020-mixed,
title = "A Mixed Learning Objective for Neural Machine Translation",
author = "Lu, Wenjie and
Zhou, Leiying and
Liu, Gongshen and
Zhang, Quanhai",
editor = "Sun, Maosong and
Li, Sujian and
Zhang, Yue and
Liu, Yang",
booktitle = "Proceedings of the 19th Chinese National Conference on Computational Linguistics",
month = oct,
year = "2020",
address = "Haikou, China",
publisher = "Chinese Information Processing Society of China",
url = "https://aclanthology.org/2020.ccl-1.90",
pages = "974--983",
abstract = "Evaluation discrepancy and overcorrection phenomenon are two common problems in neural machine translation (NMT). NMT models are generally trained with word-level learning objective, but evaluated by sentence-level metrics. Moreover, the cross-entropy loss function discourages model to generate synonymous predictions and overcorrect them to ground truth words. To address these two drawbacks, we adopt multi-task learning and propose a mixed learning objective (MLO) which combines the strength of word-level and sentence-level evaluation without modifying model structure. At word-level, it calculates semantic similarity between predicted and ground truth words. At sentence-level, it computes probabilistic n-gram matching scores of generated translations. We also combine a loss-sensitive scheduled sampling decoding strategy with MLO to explore its extensibility. Experimental results on IWSLT 2016 German-English and WMT 2019 English-Chinese datasets demonstrate that our methodology can significantly promote translation quality. The ablation study shows that both word-level and sentence-level learning objectives can improve BLEU scores. Furthermore, MLO is consistent with state-of-the-art scheduled sampling methods and can achieve further promotion.",
language = "English",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lu-etal-2020-mixed">
<titleInfo>
<title>A Mixed Learning Objective for Neural Machine Translation</title>
</titleInfo>
<name type="personal">
<namePart type="given">Wenjie</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leiying</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gongshen</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Quanhai</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2020-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<language>
<languageTerm type="text">English</languageTerm>
<languageTerm type="code" authority="iso639-2b">eng</languageTerm>
</language>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 19th Chinese National Conference on Computational Linguistics</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maosong</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sujian</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Chinese Information Processing Society of China</publisher>
<place>
<placeTerm type="text">Haikou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Evaluation discrepancy and overcorrection phenomenon are two common problems in neural machine translation (NMT). NMT models are generally trained with word-level learning objective, but evaluated by sentence-level metrics. Moreover, the cross-entropy loss function discourages model to generate synonymous predictions and overcorrect them to ground truth words. To address these two drawbacks, we adopt multi-task learning and propose a mixed learning objective (MLO) which combines the strength of word-level and sentence-level evaluation without modifying model structure. At word-level, it calculates semantic similarity between predicted and ground truth words. At sentence-level, it computes probabilistic n-gram matching scores of generated translations. We also combine a loss-sensitive scheduled sampling decoding strategy with MLO to explore its extensibility. Experimental results on IWSLT 2016 German-English and WMT 2019 English-Chinese datasets demonstrate that our methodology can significantly promote translation quality. The ablation study shows that both word-level and sentence-level learning objectives can improve BLEU scores. Furthermore, MLO is consistent with state-of-the-art scheduled sampling methods and can achieve further promotion.</abstract>
<identifier type="citekey">lu-etal-2020-mixed</identifier>
<location>
<url>https://aclanthology.org/2020.ccl-1.90</url>
</location>
<part>
<date>2020-10</date>
<extent unit="page">
<start>974</start>
<end>983</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T A Mixed Learning Objective for Neural Machine Translation
%A Lu, Wenjie
%A Zhou, Leiying
%A Liu, Gongshen
%A Zhang, Quanhai
%Y Sun, Maosong
%Y Li, Sujian
%Y Zhang, Yue
%Y Liu, Yang
%S Proceedings of the 19th Chinese National Conference on Computational Linguistics
%D 2020
%8 October
%I Chinese Information Processing Society of China
%C Haikou, China
%G English
%F lu-etal-2020-mixed
%X Evaluation discrepancy and overcorrection phenomenon are two common problems in neural machine translation (NMT). NMT models are generally trained with word-level learning objective, but evaluated by sentence-level metrics. Moreover, the cross-entropy loss function discourages model to generate synonymous predictions and overcorrect them to ground truth words. To address these two drawbacks, we adopt multi-task learning and propose a mixed learning objective (MLO) which combines the strength of word-level and sentence-level evaluation without modifying model structure. At word-level, it calculates semantic similarity between predicted and ground truth words. At sentence-level, it computes probabilistic n-gram matching scores of generated translations. We also combine a loss-sensitive scheduled sampling decoding strategy with MLO to explore its extensibility. Experimental results on IWSLT 2016 German-English and WMT 2019 English-Chinese datasets demonstrate that our methodology can significantly promote translation quality. The ablation study shows that both word-level and sentence-level learning objectives can improve BLEU scores. Furthermore, MLO is consistent with state-of-the-art scheduled sampling methods and can achieve further promotion.
%U https://aclanthology.org/2020.ccl-1.90
%P 974-983
Markdown (Informal)
[A Mixed Learning Objective for Neural Machine Translation](https://aclanthology.org/2020.ccl-1.90) (Lu et al., CCL 2020)
ACL
- Wenjie Lu, Leiying Zhou, Gongshen Liu, and Quanhai Zhang. 2020. A Mixed Learning Objective for Neural Machine Translation. In Proceedings of the 19th Chinese National Conference on Computational Linguistics, pages 974–983, Haikou, China. Chinese Information Processing Society of China.