{"id":1181840,"date":"2026-08-16T12:21:55","date_gmt":"2026-08-16T19:21:55","guid":{"rendered":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/publication\/exploiting-domain-specific-parallel-data-on-multilingual-language-models-for-low-resource-language-translation\/"},"modified":"2026-08-20T11:40:12","modified_gmt":"2026-08-20T18:40:12","slug":"exploiting-domain-specific-parallel-data-on-multilingual-language-models-for-low-resource-language-translation","status":"publish","type":"msr-research-item","link":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/publication\/exploiting-domain-specific-parallel-data-on-multilingual-language-models-for-low-resource-language-translation\/","title":{"rendered":"Exploiting Domain-Specific Parallel Data on Multilingual Language Models for Low-Resource Language Translation"},"content":{"rendered":"\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Neural Machine Translation (NMT) systems built on multilingual sequence-to-sequence Language Models (msLMs) fail to deliver expected results when the amount of parallel data for a language, as well as the language\u2019s representation in the model are limited. This restricts the capabilities of domain-specific NMT systems for low-resource languages (LRLs). As a solution, parallel data from auxiliary domains can be used either to fine-tune or to further pre-train the msLM. We present an evaluation of the effectiveness of these two techniques in the context of domain-specific LRL-NMT. We also explore the impact of domain divergence on NMT model performance. We recommend several strategies for utilizing auxiliary parallel data in building domain-specific NMT models for LRLs.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Neural Machine Translation (NMT) systems built on multilingual sequence-to-sequence Language Models (msLMs) fail to deliver expected results when the amount of parallel data for a language, as well as the language\u2019s representation in the model are limited. This restricts the capabilities of domain-specific NMT systems for low-resource languages (LRLs). As a solution, parallel data from [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":[{"type":"text","value":"Surangika Ranathungaa","user_id":0},{"type":"text","value":"Shravan Nayak","user_id":0},{"type":"text","value":"Shih-Ting Cindy Huang","user_id":0},{"type":"text","value":"Yanke Mao","user_id":0},{"type":"text","value":"Tong Su","user_id":0},{"type":"text","value":"Yun-Hsiang Ray Chan","user_id":0},{"type":"text","value":"Songchen Yuan","user_id":0},{"type":"text","value":"Anthony Rinaldi","user_id":0},{"type":"text","value":"Annie Lee","user_id":0}],"msr_publishername":"","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"ACM Trans. Asian Low Resour. Lang. Inf. Process.","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"ACM Transactions on Asian and Low-Resource Language Information Processing","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"1","msr_page_range_end":"34","msr_series":"","msr_volume":"25","msr_copyright":"","msr_conference_name":"","msr_doi":"10.1145\/3800681","msr_arxiv_id":"2412.19522","msr_mag_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_release_tracker_id":"","msr_highlight_type":"","msr_date_display_format":"","msr_main_download_label":"","msr_external_link_label":"","msr_doi_label":"","msr_published_date":"2024-12-27","msr_startdate":"","msr_presentation_date":"","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_year":2024,"msr_month":12,"msr_day":27,"msr_microsoftintellectualproperty":false,"msr_pub_id":"dfebf5772c4ed8bce39aaa236e89a97f39e0008c","msr_publication_uploader":[{"type":"url","title":"https:\/\/dblp.org\/rec\/journals\/corr\/abs-2412-19522.html","label_id":243109,"id":false,"viewUrl":false},{"type":"url","title":"https:\/\/arxiv.org\/abs\/2412.19522","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_original_fields_of_study":[],"msr_s2_paper_id":"dfebf5772c4ed8bce39aaa236e89a97f39e0008c","msr_s2_pdf_url":"https:\/\/doi.org\/10.1145\/3800681","msr_citation_count_updated":"","msr_citation_count":2,"msr_influential_citations":0,"msr_reference_count":50,"msr_s2_open_access":true,"msr_s2_author_ids":[],"msr_pub_ids":[{"provider":"s2","id":"dfebf5772c4ed8bce39aaa236e89a97f39e0008c"},{"provider":"doi","id":"10.1145\/3800681"},{"provider":"arxiv","id":"2412.19522"},{"provider":"dblp","id":"journals\/corr\/abs-2412-19522"},{"provider":"corpusid","id":"275119261"}],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13556,13545],"msr-publication-type":[193715],"msr-publisher":[],"msr-publication-cta":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[246691,249115,246808],"msr-conference":[],"msr-journal":[270372],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-1181840","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-human-language-technologies","msr-locale-en_us","msr-field-of-study-computer-science","msr-field-of-study-machine-translation","msr-field-of-study-natural-language-processing"],"msr_publishername":"","msr_edition":"ACM Trans. Asian Low Resour. Lang. Inf. Process.","msr_affiliation":"","msr_published_date":"2024-12-27","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"ACM Transactions on Asian and Low-Resource Language Information Processing","msr_volume":"25","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":0,"msr_main_download":"","msr_publicationurl":"","msr_doi":"10.1145\/3800681","msr_publication_uploader":[{"type":"url","title":"https:\/\/dblp.org\/rec\/journals\/corr\/abs-2412-19522.html","label_id":243109,"id":false,"viewUrl":false},{"type":"url","title":"https:\/\/arxiv.org\/abs\/2412.19522","label_id":243109,"id":false,"viewUrl":false}],"msr_related_uploader":[],"msr_citation_count":2,"msr_citation_count_updated":"","msr_s2_paper_id":"dfebf5772c4ed8bce39aaa236e89a97f39e0008c","msr_influential_citations":0,"msr_reference_count":50,"msr_arxiv_id":"2412.19522","msr_s2_author_ids":[],"msr_s2_open_access":true,"msr_s2_pdf_url":"https:\/\/doi.org\/10.1145\/3800681","msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Surangika Ranathungaa","user_id":0,"rest_url":false},{"type":"text","value":"Shravan Nayak","user_id":0,"rest_url":false},{"type":"text","value":"Shih-Ting Cindy Huang","user_id":0,"rest_url":false},{"type":"text","value":"Yanke Mao","user_id":0,"rest_url":false},{"type":"text","value":"Tong Su","user_id":0,"rest_url":false},{"type":"text","value":"Yun-Hsiang Ray Chan","user_id":0,"rest_url":false},{"type":"text","value":"Songchen Yuan","user_id":0,"rest_url":false},{"type":"text","value":"Anthony Rinaldi","user_id":0,"rest_url":false},{"type":"text","value":"Annie Lee","user_id":0,"rest_url":false}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[],"msr_project":[],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"article","related_content":[],"_links":{"self":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1181840","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":2,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1181840\/revisions"}],"predecessor-version":[{"id":1184382,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/1181840\/revisions\/1184382"}],"wp:attachment":[{"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1181840"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=1181840"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1181840"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=1181840"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=1181840"},{"taxonomy":"msr-publication-cta","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-cta?post=1181840"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=1181840"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1181840"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1181840"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=1181840"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=1181840"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=1181840"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1181840"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/new-cm-edgedigital.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=1181840"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}