Target-Side Augmentation for Document-Level Machine Translation
Document-level machine translation faces the challenge of data sparsity due to its long input length and a small amount of training data, increasing the risk of learning spurious patterns. To address this challenge, we propose a target-side augmentation method, introducing a data augmentation (DA) m...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | |
container_volume | |
creator | Bao, Guangsheng Teng, Zhiyang Zhang, Yue |
description | Document-level machine translation faces the challenge of data sparsity due
to its long input length and a small amount of training data, increasing the
risk of learning spurious patterns. To address this challenge, we propose a
target-side augmentation method, introducing a data augmentation (DA) model to
generate many potential translations for each source document. Learning on
these wider range translations, an MT model can learn a smoothed distribution,
thereby reducing the risk of data sparsity. We demonstrate that the DA model,
which estimates the posterior distribution, largely improves the MT
performance, outperforming the previous best system by 2.30 s-BLEU on News and
achieving new state-of-the-art on News and Europarl benchmarks. Our code is
available at https://github.com/baoguangsheng/target-side-augmentation. |
doi_str_mv | 10.48550/arxiv.2305.04505 |
format | Article |
fullrecord | <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2305_04505</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2305_04505</sourcerecordid><originalsourceid>FETCH-LOGICAL-a675-1cfa7adc66a701176e9f1a7f4dc185e305fe9ccd2f9b08c2724f8ecd20130bb53</originalsourceid><addsrcrecordid>eNotj0-PgjAUxHvZw0b3A-zJfoHiK1AKJ2N0_ZNg9rDcyaO8ahOETUWj315FT5OZTCbzY-xbQhCnSsEU_dVdgjACFUCsQH2yWYF-T734czXx-Xl_pLbH3nUtt53ny86cn4nI6UIN36E5uJZ44bE9NUNtzD4sNif6euuIFaufYrER-e96u5jnAhOthDQWNdYmSVCDlDqhzErUNq6NTBU9_ljKjKlDm1WQmlCHsU3p4UFGUFUqGrHJa3YgKP-9O6K_lU-SciCJ7uqVRAM</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Target-Side Augmentation for Document-Level Machine Translation</title><source>arXiv.org</source><creator>Bao, Guangsheng ; Teng, Zhiyang ; Zhang, Yue</creator><creatorcontrib>Bao, Guangsheng ; Teng, Zhiyang ; Zhang, Yue</creatorcontrib><description>Document-level machine translation faces the challenge of data sparsity due
to its long input length and a small amount of training data, increasing the
risk of learning spurious patterns. To address this challenge, we propose a
target-side augmentation method, introducing a data augmentation (DA) model to
generate many potential translations for each source document. Learning on
these wider range translations, an MT model can learn a smoothed distribution,
thereby reducing the risk of data sparsity. We demonstrate that the DA model,
which estimates the posterior distribution, largely improves the MT
performance, outperforming the previous best system by 2.30 s-BLEU on News and
achieving new state-of-the-art on News and Europarl benchmarks. Our code is
available at https://github.com/baoguangsheng/target-side-augmentation.</description><identifier>DOI: 10.48550/arxiv.2305.04505</identifier><language>eng</language><subject>Computer Science - Computation and Language</subject><creationdate>2023-05</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2305.04505$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2305.04505$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Bao, Guangsheng</creatorcontrib><creatorcontrib>Teng, Zhiyang</creatorcontrib><creatorcontrib>Zhang, Yue</creatorcontrib><title>Target-Side Augmentation for Document-Level Machine Translation</title><description>Document-level machine translation faces the challenge of data sparsity due
to its long input length and a small amount of training data, increasing the
risk of learning spurious patterns. To address this challenge, we propose a
target-side augmentation method, introducing a data augmentation (DA) model to
generate many potential translations for each source document. Learning on
these wider range translations, an MT model can learn a smoothed distribution,
thereby reducing the risk of data sparsity. We demonstrate that the DA model,
which estimates the posterior distribution, largely improves the MT
performance, outperforming the previous best system by 2.30 s-BLEU on News and
achieving new state-of-the-art on News and Europarl benchmarks. Our code is
available at https://github.com/baoguangsheng/target-side-augmentation.</description><subject>Computer Science - Computation and Language</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj0-PgjAUxHvZw0b3A-zJfoHiK1AKJ2N0_ZNg9rDcyaO8ahOETUWj315FT5OZTCbzY-xbQhCnSsEU_dVdgjACFUCsQH2yWYF-T734czXx-Xl_pLbH3nUtt53ny86cn4nI6UIN36E5uJZ44bE9NUNtzD4sNif6euuIFaufYrER-e96u5jnAhOthDQWNdYmSVCDlDqhzErUNq6NTBU9_ljKjKlDm1WQmlCHsU3p4UFGUFUqGrHJa3YgKP-9O6K_lU-SciCJ7uqVRAM</recordid><startdate>20230508</startdate><enddate>20230508</enddate><creator>Bao, Guangsheng</creator><creator>Teng, Zhiyang</creator><creator>Zhang, Yue</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20230508</creationdate><title>Target-Side Augmentation for Document-Level Machine Translation</title><author>Bao, Guangsheng ; Teng, Zhiyang ; Zhang, Yue</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a675-1cfa7adc66a701176e9f1a7f4dc185e305fe9ccd2f9b08c2724f8ecd20130bb53</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Computation and Language</topic><toplevel>online_resources</toplevel><creatorcontrib>Bao, Guangsheng</creatorcontrib><creatorcontrib>Teng, Zhiyang</creatorcontrib><creatorcontrib>Zhang, Yue</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Bao, Guangsheng</au><au>Teng, Zhiyang</au><au>Zhang, Yue</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Target-Side Augmentation for Document-Level Machine Translation</atitle><date>2023-05-08</date><risdate>2023</risdate><abstract>Document-level machine translation faces the challenge of data sparsity due
to its long input length and a small amount of training data, increasing the
risk of learning spurious patterns. To address this challenge, we propose a
target-side augmentation method, introducing a data augmentation (DA) model to
generate many potential translations for each source document. Learning on
these wider range translations, an MT model can learn a smoothed distribution,
thereby reducing the risk of data sparsity. We demonstrate that the DA model,
which estimates the posterior distribution, largely improves the MT
performance, outperforming the previous best system by 2.30 s-BLEU on News and
achieving new state-of-the-art on News and Europarl benchmarks. Our code is
available at https://github.com/baoguangsheng/target-side-augmentation.</abstract><doi>10.48550/arxiv.2305.04505</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | DOI: 10.48550/arxiv.2305.04505 |
ispartof | |
issn | |
language | eng |
recordid | cdi_arxiv_primary_2305_04505 |
source | arXiv.org |
subjects | Computer Science - Computation and Language |
title | Target-Side Augmentation for Document-Level Machine Translation |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T20%3A44%3A23IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Target-Side%20Augmentation%20for%20Document-Level%20Machine%20Translation&rft.au=Bao,%20Guangsheng&rft.date=2023-05-08&rft_id=info:doi/10.48550/arxiv.2305.04505&rft_dat=%3Carxiv_GOX%3E2305_04505%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |