Integrating HPSG (Head-driven Phrase Structure Grammar) with Neural Parsing for Bengali

Integrating HPSG (Head-driven Phrase Structure Grammar) with Neural Parsing for Bengali | 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 Integrating HPSG (Head-driven Phrase Structure Grammar) with Neural Parsing for Bengali Maneesha Rani Biswas This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8056398/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 In this paper, I aimed to develop a neural parser for Bangla based on simplified Head-driven Phrase Structure Grammar (HPSG) with neural network-based models. The initial stage in natural language processing is to break down the text into separate tokens. When the text corpus is huge, covering all words is inefficient regarding size of vocabulary. The effectiveness of a specific tokenization method varies on various factors, such as size of the dataset, the nature of the task, and the morphological complexity of the dataset. Due to the lack of existing HPSG-compliant treebanks for Bangla, we utilized syntactically annotated resources from existing Bangla corpora and modified them to align with simplified HPSG rule-based restructuring and data permutation. After that we modified a neural parser architecture originally designed for the Penn Treebank, replacing its encoder with multilingual pre-trained models such as XLM-RoBERTa and IndicBERT to better capture the syntactic and lexical entries of Bangla. We conducted experimental evaluations on the modified dataset, and the parser demonstrated promising results in both constituency and dependency parsing tasks. Our extensive experiments showed that the simplified HPSG Neural Parser achieved a new state-of-the-art for constituency parsing when using the same predicted part-of-speech (POS) tags as the self-attentive constituency parser. Additionally, it outperformed previous studies in dependency parsing with a higher Unlabeled Attachment Score (UAS). However, our parser remained lower Labeled Attachment Score (LAS) scores likely due to integrating HPSG with neural approaches for Bangla syntax parsing and underscoring the importance of linguistically informed treebank development in low-resource languages. Lastly, the research findings of this paper suggest that simplified HPSG should be given more attention to linguistic experts when developing treebanks for Bangla Natural Language Processing (BNLP). Neural parsing Bangla NLP Head Driven Phrase Structure Treebank Computational grammar Full Text Additional Declarations No competing interests reported. Supplementary Files ProcessedSentimentPOSTags.xlsx CleanedBanglaTense.ods Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 20 Jan, 2026 Reviewers invited by journal 14 Jan, 2026 Editor assigned by journal 14 Jan, 2026 Submission checks completed at journal 08 Nov, 2025 First submitted to journal 07 Nov, 2025 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-8056398","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":575078117,"identity":"70036349-443c-4a14-a9cc-ce5de50959c5","order_by":0,"name":"Maneesha Rani Biswas","email":"data:image/png;base64,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","orcid":"","institution":"University of Dhaka Dhaka","correspondingAuthor":true,"prefix":"","firstName":"Maneesha","middleName":"Rani","lastName":"Biswas","suffix":""}],"badges":[],"createdAt":"2025-11-07 11:08:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8056398/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8056398/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100420423,"identity":"d4a7b3e3-14dc-4589-a38c-0220a158a1a8","added_by":"auto","created_at":"2026-01-16 13:28:32","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1764117,"visible":true,"origin":"","legend":"","description":"","filename":"PaperonHPSG.docx","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/cd6dac5e7d64bd0046e905c7.docx"},{"id":100420432,"identity":"0c6c9859-397a-455c-9ff0-ef506b72cf95","added_by":"auto","created_at":"2026-01-16 13:28:35","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5101,"visible":true,"origin":"","legend":"","description":"","filename":"a0c3b37d0b0647128ed219bd38d31b2a.json","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/06b1920dddcf4f6fb09ee137.json"},{"id":100420344,"identity":"79d122ec-8464-4bda-a7f6-ec38c4447e57","added_by":"auto","created_at":"2026-01-16 13:28:19","extension":"ods","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":795412,"visible":true,"origin":"","legend":"","description":"","filename":"CleanedBanglaTense.ods","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/1e20d79f8c8e762f52b7caf5.ods"},{"id":100420315,"identity":"92579123-d4dd-445b-a1fb-5e07067dd94d","added_by":"auto","created_at":"2026-01-16 13:28:10","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31015,"visible":true,"origin":"","legend":"","description":"","filename":"ProcessedSentimentPOSTags.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/28200595fd8bcae30ee67dec.xlsx"},{"id":100420370,"identity":"23414a9d-426e-45c7-9777-fcdc7ed483d1","added_by":"auto","created_at":"2026-01-16 13:28:25","extension":"xml","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":134546,"visible":true,"origin":"","legend":"","description":"","filename":"a0c3b37d0b0647128ed219bd38d31b2a1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/c21efbef56b8f1c97c53b4ab.xml"},{"id":100420312,"identity":"2b499e03-e044-4141-8e7e-1fd773aa54b6","added_by":"auto","created_at":"2026-01-16 13:28:09","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":427663,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/f06b5b466b528a3732d2fcce.png"},{"id":100420300,"identity":"f8eeb756-99f4-4746-ab5a-3c4503b4abf3","added_by":"auto","created_at":"2026-01-16 13:28:06","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":85131,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/1666bf6bab2f5eb9fd0db6ac.jpeg"},{"id":100420374,"identity":"d2362ee7-fc81-40d0-b799-9985dc27efe2","added_by":"auto","created_at":"2026-01-16 13:28:26","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":38942,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/2abe70c34a16c10626a79125.jpeg"},{"id":100420318,"identity":"dc78d858-db11-4c03-bc14-ccaa5a5ec5e6","added_by":"auto","created_at":"2026-01-16 13:28:10","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":557215,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/280c4f3920fdde3fd43afc42.png"},{"id":100420345,"identity":"50bf311a-94b1-4cc7-a9b4-cb8674fc9201","added_by":"auto","created_at":"2026-01-16 13:28:19","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15273,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/417bbc8eef6e96e3d52cb01a.jpeg"},{"id":100420323,"identity":"15a6b2f5-440e-4ce9-a886-3e13809a206b","added_by":"auto","created_at":"2026-01-16 13:28:10","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":18247,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/935be755e943ac2117079ec1.jpeg"},{"id":100420133,"identity":"1cf4a511-0137-4574-82cd-32c85d362247","added_by":"auto","created_at":"2026-01-16 13:27:42","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":639884,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/8144e6249d4a7781810dfada.png"},{"id":100420473,"identity":"b350bdff-3918-474f-94ce-ebe778983049","added_by":"auto","created_at":"2026-01-16 13:28:46","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22986,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/80356a598c3771d89b7a497d.png"},{"id":100420376,"identity":"befd0785-ebb2-46ef-8971-ecbd8b1f636c","added_by":"auto","created_at":"2026-01-16 13:28:27","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17276,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/023bcf2f4c7cf293d007e830.png"},{"id":100420434,"identity":"3fc011cc-4ed2-471d-95f5-fe6de991c947","added_by":"auto","created_at":"2026-01-16 13:28:36","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22592,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/fef3b8138efc0945aa199aa3.png"},{"id":100420263,"identity":"def77642-1e88-4917-a3ba-efb509e58724","added_by":"auto","created_at":"2026-01-16 13:27:56","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20241,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/64d5a752b655d3b973356928.png"},{"id":100420575,"identity":"76da9ba1-175b-4336-951e-853d453c4a77","added_by":"auto","created_at":"2026-01-16 13:28:55","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14480,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/095fb6e5040f00021ced5534.png"},{"id":100420471,"identity":"105fcf9d-c371-4374-9091-02f8a2de469b","added_by":"auto","created_at":"2026-01-16 13:28:45","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17783,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/284a9f86196a177d527d4aaa.png"},{"id":100420379,"identity":"cb403104-e36d-42eb-a495-1ee32a595ea4","added_by":"auto","created_at":"2026-01-16 13:28:27","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":53436,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/483be12a09f84630dd07094d.png"},{"id":100420270,"identity":"a2c9e2de-cf0f-4f80-864e-7020071d8c21","added_by":"auto","created_at":"2026-01-16 13:28:03","extension":"xml","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128812,"visible":true,"origin":"","legend":"","description":"","filename":"a0c3b37d0b0647128ed219bd38d31b2a1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/1fff4010e57812aa6b6e38de.xml"},{"id":100420428,"identity":"4ca28594-991c-4b5f-aca9-bea5c9fb5b6c","added_by":"auto","created_at":"2026-01-16 13:28:34","extension":"html","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":144223,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/70dee8c7c46bbdf27120eed9.html"},{"id":100420849,"identity":"f80a125a-de45-4836-9b2a-6f47c162237e","added_by":"auto","created_at":"2026-01-16 13:30:05","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":915872,"visible":true,"origin":"","legend":"","description":"","filename":"PaperonHPSG.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1_covered_e01624b0-fa6e-41c4-82c0-c4450140f8e8.pdf"},{"id":100420365,"identity":"c8bd324b-ca45-44df-8b46-f675615821f7","added_by":"auto","created_at":"2026-01-16 13:28:23","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":31015,"visible":true,"origin":"","legend":"","description":"","filename":"ProcessedSentimentPOSTags.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/9f1769526bdc912d925c0e65.xlsx"},{"id":100420430,"identity":"47cb8fde-1c18-4017-90f4-f24a15ce8a4d","added_by":"auto","created_at":"2026-01-16 13:28:34","extension":"ods","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":795412,"visible":true,"origin":"","legend":"","description":"","filename":"CleanedBanglaTense.ods","url":"https://assets-eu.researchsquare.com/files/rs-8056398/v1/1ae43dafb36aceffc3711cd4.ods"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eIntegrating HPSG (Head-driven Phrase Structure Grammar) with Neural Parsing for Bengali\u003c/p\u003e","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":"language-resources-and-evaluation","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"lrev","sideBox":"Learn more about [Language Resources and Evaluation](http://link.springer.com/journal/10579)","snPcode":"10579","submissionUrl":"https://submission.nature.com/new-submission/10579/3","title":"Language Resources and Evaluation","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Neural parsing, Bangla NLP, Head Driven Phrase Structure, Treebank, Computational grammar","lastPublishedDoi":"10.21203/rs.3.rs-8056398/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8056398/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn this paper, I aimed to develop a neural parser for Bangla based on simplified Head-driven Phrase Structure Grammar (HPSG) with neural network-based models. The initial stage in natural language processing is to break down the text into separate tokens. When the text corpus is huge, covering all words is inefficient regarding size of vocabulary. The effectiveness of a specific tokenization method varies on various factors, such as size of the dataset, the nature of the task, and the morphological complexity of the dataset. Due to the lack of existing HPSG-compliant treebanks for Bangla, we utilized syntactically annotated resources from existing Bangla corpora and modified them to align with simplified HPSG rule-based restructuring and data permutation. After that we modified a neural parser architecture originally designed for the Penn Treebank, replacing its encoder with multilingual pre-trained models such as XLM-RoBERTa and IndicBERT to better capture the syntactic and lexical entries of Bangla. We conducted experimental evaluations on the modified dataset, and the parser demonstrated promising results in both constituency and dependency parsing tasks. Our extensive experiments showed that the simplified HPSG Neural Parser achieved a new state-of-the-art for constituency parsing when using the same predicted part-of-speech (POS) tags as the self-attentive constituency parser. Additionally, it outperformed previous studies in dependency parsing with a higher Unlabeled Attachment Score (UAS). However, our parser remained lower Labeled Attachment Score (LAS) scores likely due to integrating HPSG with neural approaches for Bangla syntax parsing and underscoring the importance of linguistically informed treebank development in low-resource languages. Lastly, the research findings of this paper suggest that simplified HPSG should be given more attention to linguistic experts when developing treebanks for Bangla Natural Language Processing (BNLP).\u003c/p\u003e","manuscriptTitle":"Integrating HPSG (Head-driven Phrase Structure Grammar) with Neural Parsing for Bengali","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-16 13:19:29","doi":"10.21203/rs.3.rs-8056398/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"291586388748891801353640249677489973532","date":"2026-01-20T12:04:42+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-14T13:35:54+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-14T11:55:10+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-08T14:29:35+00:00","index":"","fulltext":""},{"type":"submitted","content":"Language Resources and Evaluation","date":"2025-11-07T11:04:54+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"language-resources-and-evaluation","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"lrev","sideBox":"Learn more about [Language Resources and Evaluation](http://link.springer.com/journal/10579)","snPcode":"10579","submissionUrl":"https://submission.nature.com/new-submission/10579/3","title":"Language Resources and Evaluation","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"fd510954-1f92-43e8-9dd7-885df70b6fdf","owner":[],"postedDate":"January 16th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-01-16T13:19:29+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-16 13:19:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8056398","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8056398","identity":"rs-8056398","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}