SafeSpeech: A Three-Module Pipeline for Hate Intensity Mitigation of Social Media Texts in Indic Languages

SafeSpeech: A Three-Module Pipeline for Hate Intensity Mitigation of Social Media Texts in Indic Languages | 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 SafeSpeech: A Three-Module Pipeline for Hate Intensity Mitigation of Social Media Texts in Indic Languages Koyel Ghosh, Neeraj Kumar Singh, Joy Mahapatra, Saptarshi Saha, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5010240/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 12 Feb, 2025 Read the published version in Social Network Analysis and Mining → Version 1 posted 15 You are reading this latest preprint version Abstract Identifying and mitigating hateful, abusive, offensive comments on social media is a crucial, paramount task. It's challenging to entirely prevent such hateful content and impose rigorous censorship on social platforms while safeguarding free speech. Recent studies have focused on detecting hate speech, whereas mitigating the intensity of hate remains unexplored or somewhat complex. This paper introduces a cost-effective, straightforward, and novel three-module pipeline, SafeSpeech, for Hate Speech Classification (HSC), Hate Intensity Identification (HII), and Hate Intensity Mitigation (HIM) on social media texts. The initial module classifies text as either containing or not containing hate speech. Following this, the second module quantifies the intensity of hate associated with individual words within the classified hate speech. Lastly, the third module seeks to diminish the overall hatefulness conveyed in the text. A comprehensive experiment has been conducted using publicly available datasets in five Indic languages (Hindi, Marathi, Tamil, Telugu, and Bengali). The system undergoes thorough evaluation to assess its performance and analyze it in-depth using various automated metrics. Recognizing the limitations of automated metrics in mitigating hate speech, we augment our experiments with human evaluation, where three domain experts independently participated. BERTScore for final generated hate-mitigated texts and first classified hate texts across all languages consistently range between 0.96 and 0.99. hate speech detection pre-trained language model hate intensity identification explainability integrated gradients hate intensity mitigation Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 12 Feb, 2025 Read the published version in Social Network Analysis and Mining → Version 1 posted Editorial decision: Revision requested 18 Oct, 2024 Reviews received at journal 17 Oct, 2024 Reviews received at journal 16 Oct, 2024 Reviewers agreed at journal 10 Oct, 2024 Reviewers agreed at journal 08 Oct, 2024 Reviews received at journal 05 Oct, 2024 Reviewers agreed at journal 26 Sep, 2024 Reviewers agreed at journal 26 Sep, 2024 Reviewers agreed at journal 25 Sep, 2024 Reviewers agreed at journal 25 Sep, 2024 Reviewers agreed at journal 13 Sep, 2024 Reviewers invited by journal 13 Sep, 2024 Editor assigned by journal 04 Sep, 2024 Submission checks completed at journal 02 Sep, 2024 First submitted to journal 31 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-5010240","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":359840757,"identity":"bb10c62f-5f54-4ffa-97bd-a9bb9cc8e33f","order_by":0,"name":"Koyel Ghosh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+klEQVRIiWNgGAWjYDACCQYGZgaGBCCLufHBByDFxk68FsZmwxkgLcwkaGmT5gGJENIiH918+HMBQ5qcwfHGZmObX9vk+ZgZGD98zMGtxfDOsTTpGQw5xgZnDjY+zu27bdjGzMAsOXMbHi0zcsyYeRgqEjfcSGw2zu25zQjUwsbMi1+L8WewlvsP26Qte27bE9QiL5FjAPR1DtAWoPcZftxOJKjFQCItTZrHIM1Y8kxis2Fvw+3kNmbGZrx+kZ+RfPgzT0WyHN/xwwcf/Phz23Z+e/PBDx/x2XIATDIwKIAYjG0gHmMDbvUgWxpQGH/wKh4Fo2AUjIIRCgA1flD27VAkDAAAAABJRU5ErkJggg==","orcid":"","institution":"Central Institute of Technology Kokrajhar","correspondingAuthor":true,"prefix":"","firstName":"Koyel","middleName":"","lastName":"Ghosh","suffix":""},{"id":359840758,"identity":"5d01802d-a5a1-4999-aa76-ab8d0ed52a50","order_by":1,"name":"Neeraj Kumar Singh","email":"","orcid":"","institution":"Indian Statistical Institute Kolkata","correspondingAuthor":false,"prefix":"","firstName":"Neeraj","middleName":"Kumar","lastName":"Singh","suffix":""},{"id":359840759,"identity":"0809dc6c-560c-4c27-86da-8c74a4e6bf23","order_by":2,"name":"Joy Mahapatra","email":"","orcid":"","institution":"Indian Statistical Institute Kolkata","correspondingAuthor":false,"prefix":"","firstName":"Joy","middleName":"","lastName":"Mahapatra","suffix":""},{"id":359840760,"identity":"e5a8a4a0-a077-43b4-9f44-33e7a28afc11","order_by":3,"name":"Saptarshi Saha","email":"","orcid":"","institution":"Indian Statistical Institute Kolkata","correspondingAuthor":false,"prefix":"","firstName":"Saptarshi","middleName":"","lastName":"Saha","suffix":""},{"id":359840761,"identity":"5a2bfc2f-c730-4e53-bfcc-0c6bb5c8392a","order_by":4,"name":"Apurbalal Senapati","email":"","orcid":"","institution":"Central Institute of Technology Kokrajhar","correspondingAuthor":false,"prefix":"","firstName":"Apurbalal","middleName":"","lastName":"Senapati","suffix":""},{"id":359840762,"identity":"259cb751-1e26-4b13-b210-085381d27487","order_by":5,"name":"Utpal Garain","email":"","orcid":"","institution":"Indian Statistical Institute Kolkata","correspondingAuthor":false,"prefix":"","firstName":"Utpal","middleName":"","lastName":"Garain","suffix":""}],"badges":[],"createdAt":"2024-08-31 18:31:19","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5010240/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5010240/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s13278-024-01393-9","type":"published","date":"2025-02-12T15:57:04+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":76487441,"identity":"393bb749-b9c7-4af1-a556-7af72f635b7f","added_by":"auto","created_at":"2025-02-17 16:06:20","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":947362,"visible":true,"origin":"","legend":"","description":"","filename":"SafeSpeechfinal.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5010240/v1_covered_69313ccf-b98a-46b7-88a2-7433d34afca8.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"SafeSpeech: A Three-Module Pipeline for Hate Intensity Mitigation of Social Media Texts in Indic Languages","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":"social-network-analysis-and-mining","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"snam","sideBox":"Learn more about [Social Network Analysis and Mining](http://link.springer.com/journal/13278)","snPcode":"13278","submissionUrl":"https://submission.nature.com/new-submission/13278/3","title":"Social Network Analysis and Mining","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"hate speech detection, pre-trained language model, hate intensity identification, explainability, integrated gradients, hate intensity mitigation","lastPublishedDoi":"10.21203/rs.3.rs-5010240/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5010240/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Identifying and mitigating hateful, abusive, offensive comments on social media is a crucial, paramount task. It's challenging to entirely prevent such hateful content and impose rigorous censorship on social platforms while safeguarding free speech. Recent studies have focused on detecting hate speech, whereas mitigating the intensity of hate remains unexplored or somewhat complex. This paper introduces a cost-effective, straightforward, and novel three-module pipeline, SafeSpeech, for Hate Speech Classification (HSC), Hate Intensity Identification (HII), and Hate Intensity Mitigation (HIM) on social media texts. The initial module classifies text as either containing or not containing hate speech. Following this, the second module quantifies the intensity of hate associated with individual words within the classified hate speech. Lastly, the third module seeks to diminish the overall hatefulness conveyed in the text. A comprehensive experiment has been conducted using publicly available datasets in five Indic languages (Hindi, Marathi, Tamil, Telugu, and Bengali). The system undergoes thorough evaluation to assess its performance and analyze it in-depth using various automated metrics. Recognizing the limitations of automated metrics in mitigating hate speech, we augment our experiments with human evaluation, where three domain experts independently participated. BERTScore for final generated hate-mitigated texts and first classified hate texts across all languages consistently range between 0.96 and 0.99.","manuscriptTitle":"SafeSpeech: A Three-Module Pipeline for Hate Intensity Mitigation of Social Media Texts in Indic Languages","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-10-03 04:25:31","doi":"10.21203/rs.3.rs-5010240/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-10-18T07:18:50+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-17T21:15:00+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-16T11:04:39+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"168873595303783059143725656946897879490","date":"2024-10-10T14:55:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"278621534473679693316009417605163056458","date":"2024-10-08T18:01:33+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-10-05T07:41:20+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"287514829356330364233013445675793587177","date":"2024-09-26T10:27:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"59192868106473879006403419299407117214","date":"2024-09-26T08:48:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"284562819104552639237151573001506344219","date":"2024-09-25T14:10:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"146569612083986608171445856234156553163","date":"2024-09-25T14:05:20+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"123078777160460381034475430825174772128","date":"2024-09-13T07:44:42+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-09-13T07:19:11+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-09-04T17:18:47+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-09-02T08:27:48+00:00","index":"","fulltext":""},{"type":"submitted","content":"Social Network Analysis and Mining","date":"2024-08-31T18:30:01+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"social-network-analysis-and-mining","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"snam","sideBox":"Learn more about [Social Network Analysis and Mining](http://link.springer.com/journal/13278)","snPcode":"13278","submissionUrl":"https://submission.nature.com/new-submission/13278/3","title":"Social Network Analysis and Mining","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"67632cb0-3682-4ffd-b0ba-706d88033fd5","owner":[],"postedDate":"October 3rd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-02-17T15:59:20+00:00","versionOfRecord":{"articleIdentity":"rs-5010240","link":"https://doi.org/10.1007/s13278-024-01393-9","journal":{"identity":"social-network-analysis-and-mining","isVorOnly":false,"title":"Social Network Analysis and Mining"},"publishedOn":"2025-02-12 15:57:04","publishedOnDateReadable":"February 12th, 2025"},"versionCreatedAt":"2024-10-03 04:25:31","video":"","vorDoi":"10.1007/s13278-024-01393-9","vorDoiUrl":"https://doi.org/10.1007/s13278-024-01393-9","workflowStages":[]},"version":"v1","identity":"rs-5010240","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5010240","identity":"rs-5010240","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}