Coreference Resolution for Amharic Text using Bidirectional Encoder Representation from Transformer (BERT)

Coreference Resolution for Amharic Text using Bidirectional Encoder Representation from Transformer (BERT) | 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 Coreference Resolution for Amharic Text using Bidirectional Encoder Representation from Transformer (BERT) Lingerew Bantie, Yaregal Assabie This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7969147/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract Coreference resolution is the process of finding an entity which is refers to the same entity in a text. In coreference resolution similar entities are mentions. The task of coreference resolution is clustering all similar mentions in a text based on the index of a word. Coreference resolution is used for several NLP applications like machine translation, information extraction, name entity recognition, question answering and others to increase their effectiveness. In this work, we have proposed coreference resolution for Amharic text using bidirectional encoder representation from transformer (BERT). This method is a contextual language model that generates the semantic vectors dynamically according to the context of the words. The proposed system model has training and testing phase. The training phase includes preprocessing (cleaning, tokenization and sentence segmentation), word embedding, feature extraction and coref model. Like training phase, testing phase has its own step such as preprocessing (cleaning, tokenization and sentence segmentation) and coreference resolution as well as Amharic predicted mention. The use of word embedding in the proposed model is that it represent each word into a low dimension vector. It is a feature learning technique to obtain new features across domains for coreference resolution in Amharic text. Necessary informations are extracted from word embedding and processed data as well as Amharic characters. After we extract important features from training data we build a coreference model. Moreover, in the model bidirectional encoder representation from transformer is used to obtain basic features from embedding layer by extracting various information from both the left and right direction of the given word. To evaluate the proposed model, we conduct the experiment using Amharic dataset, which is prepared from various reliable sources for this study. The commonly used evaluation metrics for coreference resolution task are MUC, B3, CEAF-m, CEAF-e and BLANC. Experimental result demonstrate that the proposed model outperformed state-of-the-art Amharic model achieving 80%, 85.71%, 90.9%, 88.86% and 81.7% F-measure values respectively on the Amharic dataset. Physical sciences/Engineering Physical sciences/Mathematics and computing Amharic coreference resolution mention Bidirectional encoder representation from transformer Transformer NLP Coreference word embedding Full Text Additional Declarations No competing interests reported. Supplementary Files aml.coref Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 01 Feb, 2026 Reviews received at journal 29 Jan, 2026 Reviewers agreed at journal 29 Jan, 2026 Reviews received at journal 20 Jan, 2026 Reviewers agreed at journal 12 Jan, 2026 Reviewers agreed at journal 11 Jan, 2026 Reviewers invited by journal 08 Jan, 2026 Editor assigned by journal 24 Nov, 2025 Editor invited by journal 14 Nov, 2025 Submission checks completed at journal 13 Nov, 2025 First submitted to journal 13 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-7969147","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":571478367,"identity":"9e67c08a-6d07-4b6d-acfe-5b4610c06261","order_by":0,"name":"Lingerew Bantie","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEUlEQVRIiWNgGAWjYBACPmYowwBEJDDYMG4AMXjwaGFD0sLYkMCQhtCCSxsbA7IWBobDRGhh5z384uMem8Tt7GePP3hQcV52u0QC44O3bQxy9jgdxpdmOeNZWuLOnrzEhoQzt413zkhgNpzbxmCM0xZmHjNjngOHczccyDFsSGy7nbjhRgKbNG8bQ2IPPi1/DvzP3XD+DVDLv3MgLey/gVrq8Wgxfsxw4EDuhhsgWxoOgG1hBmpJwOcwxp4DyfU7Z7wxnJFwLNl4Z8/DZsk55yQMew5g18LPf8b4w48Ddsbm/DkGH3/U2MluZ08++OFNmY08ewMOa4AWSaAJgOKHAV0QBTB/wCc7CkbBKBgFo4ABAD92XWvmF1B4AAAAAElFTkSuQmCC","orcid":"","institution":"Debre Berhan University","correspondingAuthor":true,"prefix":"","firstName":"Lingerew","middleName":"","lastName":"Bantie","suffix":""},{"id":571478370,"identity":"0495dc27-8cf1-44c2-8e63-5e3016ffff80","order_by":1,"name":"Yaregal Assabie","email":"","orcid":"","institution":"Addis Ababa University","correspondingAuthor":false,"prefix":"","firstName":"Yaregal","middleName":"","lastName":"Assabie","suffix":""}],"badges":[],"createdAt":"2025-10-28 16:42:57","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7969147/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7969147/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":100008872,"identity":"738777eb-5bcd-47e1-99af-f1cf777e7a00","added_by":"auto","created_at":"2026-01-12 05:58:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":765097,"visible":true,"origin":"","legend":"","description":"","filename":"UpdatedCoreferenceResolutionforAmharictextusingBidirectionalEncoderRerpresentationfromTransformerBERTd1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7969147/v1/841cd35a80d76362023aa96a.pdf"},{"id":100008871,"identity":"3b6d2423-9a68-4e27-91d3-eb56eb173ea9","added_by":"auto","created_at":"2026-01-12 05:58:40","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5063,"visible":true,"origin":"","legend":"","description":"","filename":"af2e2a98ca664df49724c5f71507b2d3.json","url":"https://assets-eu.researchsquare.com/files/rs-7969147/v1/8d1f2a4f03981bb37535d506.json"},{"id":100008874,"identity":"86a78391-3919-4476-8b75-34887add5f64","added_by":"auto","created_at":"2026-01-12 05:58:40","extension":"coref","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":152796,"visible":true,"origin":"","legend":"","description":"","filename":"aml.coref","url":"https://assets-eu.researchsquare.com/files/rs-7969147/v1/332a48ec8100939ee0bbd011.coref"},{"id":100380853,"identity":"52b92df8-2813-4988-bfe1-8834dce1e930","added_by":"auto","created_at":"2026-01-16 10:35:53","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":373426,"visible":true,"origin":"","legend":"","description":"","filename":"UpdatedCoreferenceResolutionforAmharictextusingBidirectionalEncoderRerpresentationfromTransformerBERT.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7969147/v1_covered_78893080-3f57-4e85-ba24-fe2bc54e2765.pdf"},{"id":100361612,"identity":"5f89c5a6-ab3f-4f66-8b3a-82a41349a301","added_by":"auto","created_at":"2026-01-16 07:45:22","extension":"coref","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":152796,"visible":true,"origin":"","legend":"","description":"","filename":"aml.coref","url":"https://assets-eu.researchsquare.com/files/rs-7969147/v1/cc4b9423e1c2295231c2ca7d.coref"}],"financialInterests":"No competing interests reported.","formattedTitle":"Coreference Resolution for Amharic Text using Bidirectional Encoder Representation from Transformer (BERT)","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"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":"Amharic coreference resolution, mention, Bidirectional encoder representation from transformer, Transformer, NLP, Coreference, word embedding ","lastPublishedDoi":"10.21203/rs.3.rs-7969147/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7969147/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Coreference resolution is the process of finding an entity which is refers to the same entity in a text. In coreference resolution similar entities are mentions. The task of coreference resolution is clustering all similar mentions in a text based on the index of a word. Coreference resolution is used for several NLP applications like machine translation, information extraction, name entity recognition, question answering and others to increase their effectiveness. In this work, we have proposed coreference resolution for Amharic text using bidirectional encoder representation from transformer (BERT). This method is a contextual language model that generates the semantic vectors dynamically according to the context of the words. The proposed system model has training and testing phase. The training phase includes preprocessing (cleaning, tokenization and sentence segmentation), word embedding, feature extraction and coref model. Like training phase, testing phase has its own step such as preprocessing (cleaning, tokenization and sentence segmentation) and coreference resolution as well as Amharic predicted mention. The use of word embedding in the proposed model is that it represent each word into a low dimension vector. It is a feature learning technique to obtain new features across domains for coreference resolution in Amharic text. Necessary informations are extracted from word embedding and processed data as well as Amharic characters. After we extract important features from training data we build a coreference model. Moreover, in the model bidirectional encoder representation from transformer is used to obtain basic features from embedding layer by extracting various information from both the left and right direction of the given word. To evaluate the proposed model, we conduct the experiment using Amharic dataset, which is prepared from various reliable sources for this study. The commonly used evaluation metrics for coreference resolution task are MUC, B3, CEAF-m, CEAF-e and BLANC. Experimental result demonstrate that the proposed model outperformed state-of-the-art Amharic model achieving 80%, 85.71%, 90.9%, 88.86% and 81.7% F-measure values respectively on the Amharic dataset.","manuscriptTitle":"Coreference Resolution for Amharic Text using Bidirectional Encoder Representation from Transformer (BERT)","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-12 05:58:35","doi":"10.21203/rs.3.rs-7969147/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-02-02T03:25:49+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-29T09:48:47+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"223733668229907295113996554757680736820","date":"2026-01-29T09:36:03+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-20T11:21:14+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"263058978695088738823630161854478336606","date":"2026-01-12T11:41:06+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"103166400225994666203584937554699947412","date":"2026-01-12T00:45:57+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-08T08:38:07+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-24T10:38:18+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-11-14T13:56:42+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-13T06:05:55+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-11-13T06:04:38+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":"9dd5bccc-fb95-4d9b-a029-c4a4e2b09783","owner":[],"postedDate":"January 12th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":60813518,"name":"Physical sciences/Engineering"},{"id":60813519,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2026-03-24T06:42:24+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-12 05:58:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7969147","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7969147","identity":"rs-7969147","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}