Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions

Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions | 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 Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions Paul Fuchs, Julija Zavadlav This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9357398/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 9 You are reading this latest preprint version Abstract Foundational machine learning potentials can alleviate the accuracy and transferability limitations of classical force fields. They can substantially expedite material design and discovery by providing microscopic insights into material behavior through Molecular Dynamics simulations. However, insufficiently broad and systematically biased reference data affect the predictive quality of the learned models. These models often exhibit significant deviations from experimentally observed phase transition temperatures by several hundred kelvins. Finetuning is therefore necessary to achieve adequate accuracy in many practical problems. This work proposes a top-down finetuning strategy that corrects inaccurately predicted transition temperatures using experimental reference data. We introduce the Differentiable Transition Temperature Correction method to minimize the free energy differences between phases at the experimental target pressures and temperatures. Using a benchmark of pure Titanium at pressures up to 5 GPa, we show that our approach substantially improves the predicted phase diagram and liquid-state diffusion constant. Transition temperatures are within tens of kelvins of the experimental reference. Our model-agnostic approach can be supplemented with top-down training on additional experimental properties. In principle, it is also generally applicable to correct other free energy differences beyond the field of materials science. Thus, our approach can serve as an essential step towards highly accurate machine learning potentials. Physical sciences/Materials science Physical sciences/Physics Machine Learning Many-Body Potentials Molecular Dynamics Phase Diagram Full Text Additional Declarations No competing interests reported. Supplementary Files supplement.pdf Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 11 May, 2026 Reviews received at journal 06 May, 2026 Reviews received at journal 28 Apr, 2026 Reviewers agreed at journal 28 Apr, 2026 Reviewers agreed at journal 28 Apr, 2026 Reviewers invited by journal 28 Apr, 2026 Editor assigned by journal 10 Apr, 2026 Submission checks completed at journal 10 Apr, 2026 First submitted to journal 08 Apr, 2026 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-9357398","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":631127530,"identity":"3915517e-a9d2-4e9b-b427-b8e810999991","order_by":0,"name":"Paul Fuchs","email":"","orcid":"","institution":"Technical University of Munich","correspondingAuthor":false,"prefix":"","firstName":"Paul","middleName":"","lastName":"Fuchs","suffix":""},{"id":631127531,"identity":"1a1e9c24-faf2-4712-aa81-6625d72692c5","order_by":1,"name":"Julija Zavadlav","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA8UlEQVRIie3OsQqCQBjA8U8ObDFav5Z8hSQIB6FXuSOwIZxaGhxsseVodugh6gkyBF18AMfaC4LAUVKLqKG7teH+03fH/bgPQKX6w7SABDG+DngCGLSTISTaB6EAI0NGGvQeG8K4jJCIrWLbh0kvmp4L6iezA84pXJaCPyJWL5YCwcId2TRNPI7eTtvmMqKDDnk8Rqq7LSHdUEYqMMw8K5FW7sxoSSUh/RBwmPExstChTxIICD/XZINDK+MLm20ci+fX3XGb/ibWeprcsXQmg6SzL24lmp21tz9dfAGpNyAI+H0b/wYAZrPcTfRCpVKpVA+vTlMtzc5atwAAAABJRU5ErkJggg==","orcid":"","institution":"Technical University of Munich","correspondingAuthor":true,"prefix":"","firstName":"Julija","middleName":"","lastName":"Zavadlav","suffix":""}],"badges":[],"createdAt":"2026-04-08 13:09:01","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9357398/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9357398/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108944417,"identity":"706c24f2-0781-4e1c-b39c-7d6436b31aba","added_by":"auto","created_at":"2026-05-11 05:58:54","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":516390,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9357398/v1_covered_725429c9-7f70-4509-9df8-fcd0bd88b6a9.pdf"},{"id":108944288,"identity":"722bf45d-30e8-4af8-9a5a-f8d2e15accec","added_by":"auto","created_at":"2026-05-11 05:58:12","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":1803032,"visible":true,"origin":"","legend":"","description":"","filename":"supplement.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9357398/v1/0429ed730fb441dfda102ebe.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions","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":"npj-computational-materials","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"npjcompumats","sideBox":"Learn more about [npj Computational Materials](http://www.nature.com/npjcompumats/)","snPcode":"41524","submissionUrl":"https://mts-npjcompumats.nature.com/","title":"npj Computational Materials","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Machine Learning, Many-Body Potentials, Molecular Dynamics Phase Diagram","lastPublishedDoi":"10.21203/rs.3.rs-9357398/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9357398/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Foundational machine learning potentials can alleviate the accuracy and transferability limitations of classical force fields. They can substantially expedite material design and discovery by providing microscopic insights into material behavior through Molecular Dynamics simulations. However, insufficiently broad and systematically biased reference data affect the predictive quality of the learned models. These models often exhibit significant deviations from experimentally observed phase transition temperatures by several hundred kelvins. Finetuning is therefore necessary to achieve adequate accuracy in many practical problems. This work proposes a top-down finetuning strategy that corrects inaccurately predicted transition temperatures using experimental reference data. We introduce the Differentiable Transition Temperature Correction method to minimize the free energy differences between phases at the experimental target pressures and temperatures. Using a benchmark of pure Titanium at pressures up to 5 GPa, we show that our approach substantially improves the predicted phase diagram and liquid-state diffusion constant. Transition temperatures are within tens of kelvins of the experimental reference. Our model-agnostic approach can be supplemented with top-down training on additional experimental properties. In principle, it is also generally applicable to correct other free energy differences beyond the field of materials science. Thus, our approach can serve as an essential step towards highly accurate machine learning potentials.","manuscriptTitle":"Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-11 05:55:41","doi":"10.21203/rs.3.rs-9357398/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-11T09:12:16+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-06T09:16:25+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-28T13:47:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"35620327440765581382518397978502471006","date":"2026-04-28T11:05:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"318283158753931135977802882954038799403","date":"2026-04-28T10:07:31+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-28T07:16:47+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-10T08:46:00+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-10T04:39:24+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Computational Materials","date":"2026-04-08T13:06:05+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-computational-materials","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"npjcompumats","sideBox":"Learn more about [npj Computational Materials](http://www.nature.com/npjcompumats/)","snPcode":"41524","submissionUrl":"https://mts-npjcompumats.nature.com/","title":"npj Computational Materials","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"39d4f7ed-0323-4c3f-bffd-22ffbc82f88b","owner":[],"postedDate":"May 11th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-11T09:12:16+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-06T09:16:25+00:00","index":16,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":67174665,"name":"Physical sciences/Materials science"},{"id":67174666,"name":"Physical sciences/Physics"}],"tags":[],"updatedAt":"2026-05-11T09:26:20+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-11 05:55:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9357398","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9357398","identity":"rs-9357398","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}