MAPPING: Debiasing Graph Neural Networks for Fair Node Classification with Limited Sensitive Information Leakage

MAPPING: Debiasing Graph Neural Networks for Fair Node Classification with Limited Sensitive Information Leakage | 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 MAPPING: Debiasing Graph Neural Networks for Fair Node Classification with Limited Sensitive Information Leakage Ying Song, Balaji Palanisamy This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-3903316/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 06 Nov, 2024 Read the published version in World Wide Web → Version 1 posted 13 You are reading this latest preprint version Abstract Despite remarkable success in diverse web-based applications, Graph Neural Networks(GNNs) inherit and further exacerbate historical discrimination and social stereotypes, which critically hinder their deployments in high-stake domains such as online clinical diagnosis, financial crediting, etc. However, existing research in fair graph learning typically favors pairwise constraints to achieve fairness but fails to cast off dimensional limitations and generalize them into multiple sensitive attributes; besides, most studies focus on in-processing techniques to enforce and calibrate fairness, constructing a model-agnostic debiasing GNN framework at the pre-processing stage to prevent downstream misuses and improve training reliability is still largely under-explored. Furthermore, previous work tends to enhance either fairness or privacy individually but few probes into how fairness issues trigger privacy concerns and whether such concerns can be alleviated with fairness intervention. In this paper, we propose a novel model-agnostic debiasing framework named MAPPING (\underline{M}asking \underline{A}nd \underline{P}runing and Message-\underline{P}assing train\underline{ING}) for fair node classification, in which we adopt the distance covariance($dCov$)-based fairness constraints to simultaneously reduce feature and topology biases under multiple sensitive memberships, and combine them with adversarial debiasing to confine the risks of sensitive attribute inference. Experiments on real-world datasets with different GNN variants demonstrate the effectiveness and flexibility of MAPPING. Our results show that MAPPING can achieve better trade-offs between utility and fairness, and mitigate privacy risks of sensitive information leakage. Graph Neural Networks Group Fairness Privacy Risks Distance Covariance Adversarial Training Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 06 Nov, 2024 Read the published version in World Wide Web → Version 1 posted Editorial decision: Revision requested 31 Aug, 2024 Reviews received at journal 31 Aug, 2024 Reviews received at journal 26 Aug, 2024 Reviewers agreed at journal 18 Aug, 2024 Reviews received at journal 18 Aug, 2024 Reviewers agreed at journal 18 Aug, 2024 Reviewers agreed at journal 16 Aug, 2024 Reviews received at journal 15 Apr, 2024 Reviewers agreed at journal 29 Mar, 2024 Reviewers invited by journal 29 Mar, 2024 Editor assigned by journal 29 Jan, 2024 Submission checks completed at journal 29 Jan, 2024 First submitted to journal 27 Jan, 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-3903316","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":269952078,"identity":"f178b5f6-e390-4bcb-be5f-34193576157b","order_by":0,"name":"Ying Song","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAvUlEQVRIie3PsQrCMBCA4QuFuNTOOtVHuF3xWSoF3YogdC4InQTXPs6FDN3aB+hS8QUKLg4VvNrFQeiNDvmHhMB9kANwuf4xD4A63ID+vLSMKFMc96C1JyWMrN9ZnpYSLGct+VgnQXgl6FIrINZHWmBzGj6mikpGgBCbXc7Em+ciwh+LsBrJS0YAiZBGoiRkybuYDGPeJUZzqQ7TJKjL+6Pvt0l4Nrf2ma6nyYrGOxoOmp7nwuyLuFwul+tnb2kdOzBh2eK0AAAAAElFTkSuQmCC","orcid":"","institution":"University of Pittsburgh","correspondingAuthor":true,"prefix":"","firstName":"Ying","middleName":"","lastName":"Song","suffix":""},{"id":269952079,"identity":"cb7d64eb-3ece-41f9-aaa1-2cec6b9064cc","order_by":1,"name":"Balaji Palanisamy","email":"","orcid":"","institution":"University of Pittsburgh","correspondingAuthor":false,"prefix":"","firstName":"Balaji","middleName":"","lastName":"Palanisamy","suffix":""}],"badges":[],"createdAt":"2024-01-27 14:49:52","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-3903316/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-3903316/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s11280-024-01312-0","type":"published","date":"2024-11-06T15:57:27+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":68750133,"identity":"45f155b2-abdd-4803-931f-741282693c35","added_by":"auto","created_at":"2024-11-11 16:10:36","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":12428097,"visible":true,"origin":"","legend":"","description":"","filename":"MAPPINGManuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-3903316/v1_covered_996fc53a-cc9a-488e-a116-8cc10bef465d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"MAPPING: Debiasing Graph Neural Networks for Fair Node Classification with Limited Sensitive Information Leakage","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":"world-wide-web","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"wwwj","sideBox":"Learn more about [World Wide Web](http://link.springer.com/journal/11280)","snPcode":"11280","submissionUrl":"https://submission.nature.com/new-submission/11280/3","title":"World Wide Web","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Graph Neural Networks, Group Fairness, Privacy Risks, Distance Covariance, Adversarial Training","lastPublishedDoi":"10.21203/rs.3.rs-3903316/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-3903316/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Despite remarkable success in diverse web-based applications, Graph Neural Networks(GNNs) inherit and further exacerbate historical discrimination and social stereotypes, which critically hinder their deployments in high-stake domains such as online clinical diagnosis, financial crediting, etc. However, existing research in fair graph learning typically favors pairwise constraints to achieve fairness but fails to cast off dimensional limitations and generalize them into multiple sensitive attributes; besides, most studies focus on in-processing techniques to enforce and calibrate fairness, constructing a model-agnostic debiasing GNN framework at the pre-processing stage to prevent downstream misuses and improve training reliability is still largely under-explored. Furthermore, previous work tends to enhance either fairness or privacy individually but few probes into how fairness issues trigger privacy concerns and whether such concerns can be alleviated with fairness intervention. In this paper, we propose a novel model-agnostic debiasing framework named MAPPING (\\underline{M}asking \\underline{A}nd \\underline{P}runing and Message-\\underline{P}assing train\\underline{ING}) for fair node classification, in which we adopt the distance covariance($dCov$)-based fairness constraints to simultaneously reduce feature and topology biases under multiple sensitive memberships, and combine them with adversarial debiasing to confine the risks of sensitive attribute inference. Experiments on real-world datasets with different GNN variants demonstrate the effectiveness and flexibility of MAPPING. Our results show that MAPPING can achieve better trade-offs between utility and fairness, and mitigate privacy risks of sensitive information leakage.","manuscriptTitle":"MAPPING: Debiasing Graph Neural Networks for Fair Node Classification with Limited Sensitive Information Leakage","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-01-31 06:37:49","doi":"10.21203/rs.3.rs-3903316/v1","editorialEvents":[{"type":"communityComments","content":1},{"type":"decision","content":"Revision requested","date":"2024-09-01T00:42:11+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-31T14:30:40+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-26T12:01:29+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"33300325203990775592597422562628787434","date":"2024-08-18T11:46:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-08-18T11:21:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"61178289256565722277103046013815075166","date":"2024-08-18T11:14:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"32386264649769015502159764693498822960","date":"2024-08-16T12:19:20+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-04-16T01:54:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"140a10ef-caa5-4bdd-b808-760304487a0c","date":"2024-03-29T22:24:16+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-03-29T21:22:01+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-01-30T01:30:22+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-01-29T17:49:34+00:00","index":"","fulltext":""},{"type":"submitted","content":"World Wide Web","date":"2024-01-27T14:44:12+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"world-wide-web","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"wwwj","sideBox":"Learn more about [World Wide Web](http://link.springer.com/journal/11280)","snPcode":"11280","submissionUrl":"https://submission.nature.com/new-submission/11280/3","title":"World Wide Web","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"7d565d25-bf2d-4c7e-bae1-cdcc25a22528","owner":[],"postedDate":"January 31st, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-11-11T16:04:53+00:00","versionOfRecord":{"articleIdentity":"rs-3903316","link":"https://doi.org/10.1007/s11280-024-01312-0","journal":{"identity":"world-wide-web","isVorOnly":false,"title":"World Wide Web"},"publishedOn":"2024-11-06 15:57:27","publishedOnDateReadable":"November 6th, 2024"},"versionCreatedAt":"2024-01-31 06:37:49","video":"","vorDoi":"10.1007/s11280-024-01312-0","vorDoiUrl":"https://doi.org/10.1007/s11280-024-01312-0","workflowStages":[]},"version":"v1","identity":"rs-3903316","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-3903316","identity":"rs-3903316","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}