Enhancing Neural Machine Translation for Low-Resource Languages through Innovative Techniques and Approaches

Enhancing Neural Machine Translation for Low-Resource Languages through Innovative Techniques and Approaches | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Enhancing Neural Machine Translation for Low-Resource Languages through Innovative Techniques and Approaches Qing Yu, Zhaojie Liu, Zhenwei Xu, Yaoyong Zhou, Wushouer Silamu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4851167/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 5 You are reading this latest preprint version Abstract Machine translation is a critical area in natural language processing, aiming to achieve automatic translation between different languages. Despite significant progress in multilingual translation tasks with neural machine translation models, many challenges remain in handling low-resource languages. These challenges mainly arise from the limited availability of parallel corpora for training and the differences in grammatical structures and expressions among languages. Based on this, this study proposes an iterative ensemble pruning data augmentation strategy, combined with a multimodal model that incorporates glyph and pinyin features. First, the iterative ensemble pruning method is used to enhance data quality and quantity. Second, a multimodal model utilizing glyph and pinyin features is constructed to achieve more accurate translation of low-resource languages. Finally, these two subtasks are collaboratively optimized to improve model performance. Experimental validation on the Uighur-Chinese medical bilingual dataset provided by Xinjiang University and the CCEval multilingual machine translation evaluation set demonstrates that the proposed method outperforms existing baseline models in terms of translation accuracy and robustness. The experimental results indicate that the iterative ensemble pruning strategy effectively improves data quality, and the multimodal model significantly enhances the translation performance of low-resource languages. In summary, the proposed method not only performs excellently in low-resource language machine translation but also provides new ideas and methods for data augmentation and multimodal modeling in other natural language processing tasks, having significant research significance and application value. Biological sciences/Computational biology and bioinformatics/Computational neuroscience Biological sciences/Computational biology and bioinformatics/Programming language and code Physical sciences/Mathematics and computing/Computer science Physical sciences/Mathematics and computing/Information technology Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers invited by journal 27 Dec, 2024 Editor assigned by journal 15 Nov, 2024 Editor invited by journal 20 Aug, 2024 Submission checks completed at journal 20 Aug, 2024 First submitted to journal 02 Aug, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4851167","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":354256090,"identity":"13a552e8-eb9d-4784-a934-8835ca5d8cc2","order_by":0,"name":"Qing Yu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA70lEQVRIie3RMYvCMBTA8RcCdXmta0oPex8hxemgcF8lUuh0QydXC4XeIs76LRS/QI8MtxRdO9al80HhOA4HUxXH2FEw/+2F/PKGAJhMD1oBwNCntAB2HfuQ8CX4tER/oopD2CO/vaCNV9FGJv8SSYa/7C2HkVMJ0iZaEidytZA4oPaWuzmM3UpQb6klH1za826LvakVmawrYVHsQ0BiXSgy60fwL+4I6bYIfo+4ZZNIOw0xyKwxZzsWrMpD5umI8x1tWzyyd38oG49NQ1+dfLU68loAB5JfBsrOn0lSDQDwU0XgeBnIj/auyWQyPWsn5sVNOpawK/0AAAAASUVORK5CYII=","orcid":"","institution":"Xinjiang University","correspondingAuthor":true,"prefix":"","firstName":"Qing","middleName":"","lastName":"Yu","suffix":""},{"id":354256091,"identity":"ea47fa92-e77b-412a-9b37-390aedecd999","order_by":1,"name":"Zhaojie Liu","email":"","orcid":"","institution":"Xinjiang University","correspondingAuthor":false,"prefix":"","firstName":"Zhaojie","middleName":"","lastName":"Liu","suffix":""},{"id":354256092,"identity":"7460739a-9473-4a65-9ce5-f2584ea879d0","order_by":2,"name":"Zhenwei Xu","email":"","orcid":"","institution":"Xinjiang University","correspondingAuthor":false,"prefix":"","firstName":"Zhenwei","middleName":"","lastName":"Xu","suffix":""},{"id":354256093,"identity":"9277efec-cf20-4603-8708-22e6716dfeca","order_by":3,"name":"Yaoyong Zhou","email":"","orcid":"","institution":"Xinjiang University","correspondingAuthor":false,"prefix":"","firstName":"Yaoyong","middleName":"","lastName":"Zhou","suffix":""},{"id":354256094,"identity":"fde2936d-5bb3-4703-80a7-01cb6dc84a27","order_by":4,"name":"Wushouer Silamu","email":"","orcid":"","institution":"Xinjiang University","correspondingAuthor":false,"prefix":"","firstName":"Wushouer","middleName":"","lastName":"Silamu","suffix":""}],"badges":[],"createdAt":"2024-08-03 03:29:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4851167/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4851167/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":64655703,"identity":"5f4859ea-5d80-4f42-93cb-c35875441bc9","added_by":"auto","created_at":"2024-09-17 06:45:21","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":629237,"visible":true,"origin":"","legend":"","description":"","filename":"TemplateforsubmissionstoScientificReports102.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4851167/v1_covered_7d4271da-f287-4483-97f6-80790ec131cc.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Enhancing Neural Machine Translation for Low-Resource Languages through Innovative Techniques and Approaches","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4851167/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4851167/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Machine translation is a critical area in natural language processing, aiming to achieve automatic translation between different languages. Despite significant progress in multilingual translation tasks with neural machine translation models, many challenges remain in handling low-resource languages. These challenges mainly arise from the limited availability of parallel corpora for training and the differences in grammatical structures and expressions among languages. Based on this, this study proposes an iterative ensemble pruning data augmentation strategy, combined with a multimodal model that incorporates glyph and pinyin features. First, the iterative ensemble pruning method is used to enhance data quality and quantity. Second, a multimodal model utilizing glyph and pinyin features is constructed to achieve more accurate translation of low-resource languages. Finally, these two subtasks are collaboratively optimized to improve model performance. Experimental validation on the Uighur-Chinese medical bilingual dataset provided by Xinjiang University and the CCEval multilingual machine translation evaluation set demonstrates that the proposed method outperforms existing baseline models in terms of translation accuracy and robustness. The experimental results indicate that the iterative ensemble pruning strategy effectively improves data quality, and the multimodal model significantly enhances the translation performance of low-resource languages. In summary, the proposed method not only performs excellently in low-resource language machine translation but also provides new ideas and methods for data augmentation and multimodal modeling in other natural language processing tasks, having significant research significance and application value.","manuscriptTitle":"Enhancing Neural Machine Translation for Low-Resource Languages through Innovative Techniques and Approaches","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-09-17 06:37:14","doi":"10.21203/rs.3.rs-4851167/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewersInvited","content":"","date":"2024-12-27T07:43:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-11-15T10:57:06+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-08-20T17:12:50+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-08-20T05:44:44+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-08-03T03:26:59+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c6ce148d-bba8-4399-a289-e04a77dc5225","owner":[],"postedDate":"September 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":37647501,"name":"Biological sciences/Computational biology and bioinformatics/Computational neuroscience"},{"id":37647502,"name":"Biological sciences/Computational biology and bioinformatics/Programming language and code"},{"id":37647503,"name":"Physical sciences/Mathematics and computing/Computer science"},{"id":37647504,"name":"Physical sciences/Mathematics and computing/Information technology"}],"tags":[],"updatedAt":"2024-09-17T06:37:14+00:00","versionOfRecord":[],"versionCreatedAt":"2024-09-17 06:37:14","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4851167","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4851167","identity":"rs-4851167","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}