ViVerse-A1: Ascending Generating Vietnamese Poetry with Semantic and Cultural Fidelity

ViVerse-A1: Ascending Generating Vietnamese Poetry with Semantic and Cultural Fidelity | 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 Research Article ViVerse-A1: Ascending Generating Vietnamese Poetry with Semantic and Cultural Fidelity Phu Lam, Hung Tran, Tram Doan, Binh Nguyen, Son Huynh This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9115864/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Despite their increasing sophistication in content generation, contemporary Large Language Models (LLMs) consistently fail to master the nuanced artistry of Vietnamese poetry, particularly its intricate semantic layers, figurative language, phonaesthetics, and deep-seated cultural resonance. This research confronts this challenge with the introduction of ViVerse-A1, a bespoke neural architecture optimized for Vietnamese poetic expression. ViVerse-A1's novelty lies in its hybrid training methodology, which synergizes targeted fine-tuning with advanced auxiliary frameworks, including a Masked Language Model and our proposed Masked-Token Reasoning Language Model. This design empowers the model to develop a granular understanding of the structural, prosodic, lexical, and diverse conventions of Vietnamese literature. Rigorous comparative analyses confirm that ViVerse-A1's performance surpasses that of leading state-of-the-art models, demonstrating a unique ability to generate poetry that is both technically proficient and expressively profound. The success of ViVerse-A1 signals a promising trajectory for AI-driven systems designed for deep semantic and cultural intelligence. To facilitate further research, the source code and all related materials are provided publicly at https://github.com/tph-kds/ViVerse_A1 AI language processing Vietnamese poem generation fine-tuning ViVerse-A1 Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted 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-9115864","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":606493480,"identity":"80acaed5-1caf-4957-9f35-764984f48a81","order_by":0,"name":"Phu Lam","email":"","orcid":"","institution":"Ho Chi Minh City University of Science","correspondingAuthor":false,"prefix":"","firstName":"Phu","middleName":"","lastName":"Lam","suffix":""},{"id":606493485,"identity":"885f0840-9b6d-463a-af6d-0d464e0de237","order_by":1,"name":"Hung Tran","email":"","orcid":"","institution":"Ho Chi Minh City University of Science","correspondingAuthor":false,"prefix":"","firstName":"Hung","middleName":"","lastName":"Tran","suffix":""},{"id":606493488,"identity":"37a4187b-5ba8-4495-9659-941317c910ae","order_by":2,"name":"Tram Doan","email":"","orcid":"","institution":"Ho Chi Minh City University of Science","correspondingAuthor":false,"prefix":"","firstName":"Tram","middleName":"","lastName":"Doan","suffix":""},{"id":606493490,"identity":"f8eea757-5c88-42c4-913e-d13b5da83a8b","order_by":3,"name":"Binh Nguyen","email":"","orcid":"","institution":"Ho Chi Minh City University of Science","correspondingAuthor":false,"prefix":"","firstName":"Binh","middleName":"","lastName":"Nguyen","suffix":""},{"id":606493499,"identity":"5b424770-eee4-42f7-8c7b-d5bd58c1690d","order_by":4,"name":"Son Huynh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYBACAzBZcQDKZSNGC1jRGZK1MLaRosVcvv3h48J5d+R1Z6Q/YPhQdpjB4HYDfi2WbTzGxjO3PTPcdiPHgHHGOaCWOwfwazE4xsMmzbvtcILZjRwGZt62wwySMxIIaWF/Js07B6Ql/QHzX+K0MJhJ8zaAtCQYMDMCtfBLENSSY2zMc+yw4bYzbwwO9pxL5yGs5fDxh495ag7Lmx1Pf/jgR5m1HBshLSjgABDzkKB+FIyCUTAKRgEuAACdTkQKWAZVwgAAAABJRU5ErkJggg==","orcid":"","institution":"Ho Chi Minh City University of Science","correspondingAuthor":true,"prefix":"","firstName":"Son","middleName":"","lastName":"Huynh","suffix":""}],"badges":[],"createdAt":"2026-03-13 14:39:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9115864/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9115864/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106404175,"identity":"2180f9e0-2024-49d1-8d7e-af94bbbb2cbb","added_by":"auto","created_at":"2026-04-08 09:15:36","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2851015,"visible":true,"origin":"","legend":"","description":"","filename":"ViVerseA1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9115864/v1_covered_d9a26425-a06c-4c5d-9c4f-c6de0be4eeda.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"ViVerse-A1: Ascending Generating Vietnamese Poetry with Semantic and Cultural Fidelity","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"AI language processing, Vietnamese poem generation, fine-tuning, ViVerse-A1","lastPublishedDoi":"10.21203/rs.3.rs-9115864/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9115864/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Despite their increasing sophistication in content generation, contemporary Large Language Models (LLMs) consistently fail to master the nuanced artistry of Vietnamese poetry, particularly its intricate semantic layers, figurative language, phonaesthetics, and deep-seated cultural resonance. This research confronts this challenge with the introduction of ViVerse-A1, a bespoke neural architecture optimized for Vietnamese poetic expression. ViVerse-A1's novelty lies in its hybrid training methodology, which synergizes targeted fine-tuning with advanced auxiliary frameworks, including a Masked Language Model and our proposed Masked-Token Reasoning Language Model. This design empowers the model to develop a granular understanding of the structural, prosodic, lexical, and diverse conventions of Vietnamese literature. Rigorous comparative analyses confirm that ViVerse-A1's performance surpasses that of leading state-of-the-art models, demonstrating a unique ability to generate poetry that is both technically proficient and expressively profound. The success of ViVerse-A1 signals a promising trajectory for AI-driven systems designed for deep semantic and cultural intelligence. To facilitate further research, the source code and all related materials are provided publicly at https://github.com/tph-kds/ViVerse_A1","manuscriptTitle":"ViVerse-A1: Ascending Generating Vietnamese Poetry with Semantic and Cultural Fidelity","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-07 11:03:58","doi":"10.21203/rs.3.rs-9115864/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"23557806-a063-46d3-bb13-e7c03d04fec5","owner":[],"postedDate":"April 7th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-07T11:03:58+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-07 11:03:58","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9115864","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9115864","identity":"rs-9115864","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}