Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding

Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding | 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 Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding Cecilia Clementi, Alessandro Caruso, Jacopo Venturin, Lorenzo Giambagli, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6065047/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 18 Feb, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Abstract Graph Neural Networks (GNNs) are routinely used in molecular physics, social sciences, and economics to model many-body interactions in graph-like systems. However, GNNs are inherently local and can suffer from information flow bottlenecks. This is particularly problematic when modeling large molecular systems, where dispersion forces and local electric field variations drive collective structural changes. Existing solutions face challenges related to computational cost and scalability. We introduce RANGE, a model-agnostic framework that employs an attention-based aggregation-broadcast mechanism that significantly reduces oversquashing effects, and achieves remarkable accuracy in capturing long-range interactions at a negligible computational cost. Notably, RANGE is the first virtual-node message-passing implementation to integrate attention with positional encodings and regularization to dynamically expand virtual representations. This work lays the foundation for next-generation of machine-learned force fields, offering accurate and efficient modeling of long-range interactions for simulating large molecular systems. Physical sciences/Chemistry/Theoretical chemistry/Computational chemistry Physical sciences/Mathematics and computing/Computational science Full Text Additional Declarations There is NO Competing Interest. Supplementary Files RANGESIFinal.pdf Supplementary Information Cite Share Download PDF Status: Published Journal Publication published 18 Feb, 2026 Read the published version in Nature Communications → 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-6065047","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":428521364,"identity":"317b1867-3876-4486-832c-30a6ceb5088b","order_by":0,"name":"Cecilia Clementi","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0001-9221-2358","institution":"Freie Universität Berlin","correspondingAuthor":true,"prefix":"","firstName":"Cecilia","middleName":"","lastName":"Clementi","suffix":""},{"id":428521365,"identity":"161c686f-f582-4eec-b218-da9f0db03e52","order_by":1,"name":"Alessandro Caruso","email":"","orcid":"https://orcid.org/0000-0002-0907-0668","institution":"Freie Universität Berlin","correspondingAuthor":false,"prefix":"","firstName":"Alessandro","middleName":"","lastName":"Caruso","suffix":""},{"id":428521366,"identity":"420df889-2811-46a0-bdfa-94fcec11e217","order_by":2,"name":"Jacopo Venturin","email":"","orcid":"https://orcid.org/0009-0003-7347-3036","institution":"Freie Universität Berlin","correspondingAuthor":false,"prefix":"","firstName":"Jacopo","middleName":"","lastName":"Venturin","suffix":""},{"id":428521367,"identity":"eccdf099-e8bc-431a-a860-33223d07c968","order_by":3,"name":"Lorenzo Giambagli","email":"","orcid":"","institution":"Freie Universität Berlin","correspondingAuthor":false,"prefix":"","firstName":"Lorenzo","middleName":"","lastName":"Giambagli","suffix":""},{"id":428521368,"identity":"1df8c43c-a298-46a7-a1e0-0dacbdb661b7","order_by":4,"name":"Edoardo Rolando","email":"","orcid":"https://orcid.org/0009-0001-8276-1406","institution":"Freie Universität Berlin","correspondingAuthor":false,"prefix":"","firstName":"Edoardo","middleName":"","lastName":"Rolando","suffix":""},{"id":428521369,"identity":"210ad9e7-0248-497f-8b67-142581ceacc0","order_by":5,"name":"Frank Noé","email":"","orcid":"","institution":"Microsoft Research, AI for Science","correspondingAuthor":false,"prefix":"","firstName":"Frank","middleName":"","lastName":"Noé","suffix":""}],"badges":[],"createdAt":"2025-02-19 14:33:59","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6065047/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6065047/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41467-026-69715-3","type":"published","date":"2026-02-18T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":103052882,"identity":"48d0359c-7659-4976-8e14-c0f656749d15","added_by":"auto","created_at":"2026-02-20 08:07:56","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3032229,"visible":true,"origin":"","legend":"Article File","description":"","filename":"RANGEFinal.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6065047/v1_covered_0b2e42b6-3036-47e0-b69a-51e8dced777c.pdf"},{"id":78626663,"identity":"de1f540e-f724-42a6-8215-99108b994de1","added_by":"auto","created_at":"2025-03-17 02:16:57","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":8115025,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"RANGESIFinal.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6065047/v1/1a24ce45a562ec55b3e1c10a.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6065047/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6065047/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Graph Neural Networks (GNNs) are routinely used in molecular physics, social sciences, and economics to model many-body interactions in graph-like systems. However, GNNs are inherently local and can suffer from information flow bottlenecks. This is particularly problematic when modeling large molecular systems, where dispersion forces and local electric field variations drive collective structural changes. Existing solutions face challenges related to computational cost and scalability. We introduce RANGE, a model-agnostic framework that employs an attention-based aggregation-broadcast mechanism that significantly reduces oversquashing effects, and achieves remarkable accuracy in capturing long-range interactions at a negligible computational cost. Notably, RANGE is the first virtual-node message-passing implementation to integrate attention with positional encodings and regularization to dynamically expand virtual representations. This work lays the foundation for next-generation of machine-learned force fields, offering accurate and efficient modeling of long-range interactions for simulating large molecular systems.","manuscriptTitle":"Extending the RANGE of Graph Neural Networks: Relaying Attention Nodes for Global Encoding","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-17 02:16:52","doi":"10.21203/rs.3.rs-6065047/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"006fa1af-01e2-4d75-b539-c4b9c6c6f6c5","owner":[],"postedDate":"March 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":45656441,"name":"Physical sciences/Chemistry/Theoretical chemistry/Computational chemistry"},{"id":45656442,"name":"Physical sciences/Mathematics and computing/Computational science"}],"tags":[],"updatedAt":"2026-02-20T08:07:44+00:00","versionOfRecord":{"articleIdentity":"rs-6065047","link":"https://doi.org/10.1038/s41467-026-69715-3","journal":{"identity":"nature-communications","isVorOnly":false,"title":"Nature Communications"},"publishedOn":"2026-02-18 05:00:00","publishedOnDateReadable":"February 18th, 2026"},"versionCreatedAt":"2025-03-17 02:16:52","video":"","vorDoi":"10.1038/s41467-026-69715-3","vorDoiUrl":"https://doi.org/10.1038/s41467-026-69715-3","workflowStages":[]},"version":"v1","identity":"rs-6065047","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6065047","identity":"rs-6065047","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}