[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120847-en":3,"doc-seo-120847-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},120847,1649267921044,"Ava Thompson","https://us-avatar.wpscdn.com/avatar/1800007509477c92dfb?_k=1782875107921204101",8,"Research & Report","Dynamic heterogeneity at the experimental glass transition predicted by transferable machine learning - Research overview","A transferable machine learning model predicts structural relaxation and dynamic heterogeneity from amorphous supercooled liquid structures across wide temperature and time-scale ranges with high accuracy. By leveraging network transferability, the method reaches the experimental glass transition temperature Tg, where structural relaxation is inaccessible to molecular dynamics. The study finds that near Tg the heterogeneity’s strength, geometry, and characteristic length scale change more slowly than at higher temperatures. Results suggest ML can reveal physical aspects of the glass transition beyond conventional simulations.","Dynamic heterogeneity at the experimental glass transition predicted by transferable  \nmachine learning  \narXiv :2310 .20252v1 [ cond-mat .soft] 31 Oct 2023  \nGerhard Jung, 1 Giulio Biroli,2 and Ludovic Berthier 1, 3  \n1 Laboratoire Charles Coulomb (L2C), Universit´e de Montpellier, CNRS, 34095 Montpellier, France  \n2 Laboratoire de Physique de l’Ecole Normale Sup´erieure, ENS, Universit´e PSL,  \nCNRS, Sorbonne Universit´e, Universit´e de Paris, F-75005 Paris, France  \n3 Gulliver, UMR CNRS 7083, ESPCI Paris, PSL Research University, 75005 Paris, France  \n(Dated: November 1, 2023)  \nWe develop a transferable machine learning model which predicts structural relaxation from amorphous supercooled liquid structures. The trained networks are able to predict dynamic heterogeneity across a broad range of temperatures and time scales with excellent accuracy and transferability. We use the network transferability to predict dynamic heterogeneity down to the experimental glass transition temperature, Tg , where structural relaxation cannot be analyzed using molecular dynamics simulations. The results indicate that the strength, the geometry and the characteristic length scale of the dynamic heterogeneity evolve much more slowly near Tg compared to their evolution at higher temperatures. Our results show that machine learning techniques can provide physical insights on the nature of the glass transition that cannot be gained using conventional simulation techniques.  \nI. INTRODUCTION  \nDense liquids display a drastic slowing down of structural relaxation when approaching the experimental glass transition temperature [1, 2] . The glass transition is characterized by several important properties, such as a very homogeneous amorphous structure but a strongly heterogeneous relaxation dynamics, leading to the spatial coexistence of frozen and active regions [3] . Understanding the connection between the microstructure and dynamic heterogeneity is an important field of research [4–7] .  \nOver the years, several structural order parameters have been proposed which show some degree of correlation with the local relaxation dynamics, including density [8], potential energy [9, 10], geometry of Voronoi cells [11], soft modes [4], locally-favored structures [5, 12– 14], and more [6, 7 , 15 , 16] . Recently, the application of machine learning (ML) techniques to automatically construct suitable structural order parameters has significantly advanced this line of research. The range of methodologies includes unsupervised learning to automatically detect structural heterogeneities [8, 17–19] and supervised learning using linear regression [20–22], support vector machines [23–25], multilayer perceptrons (MLP) [26] and graph neural networks (GNN) [27–30] . The performance of these techniques significantly surpasses traditional approaches based on hand-made order parameters, and allows to infer the microscopic structural relaxation from structural properties with high accuracy, including aspects of dynamic heterogeneity [26, 28] .  \nThanks to this progress, ML approaches lead to new physical results. Applications of trained neural networks have used scalability in system size to extract new results on dynamic length scales and the geometry of rearranging domains [26], transferability to other state points to analyze structural differences between strong and fragile glass formers [31] . Trained models  \nwere also used to construct effective glass models [32] . While the performance of ML approaches is remarkable, one of the main drawbacks of the supervised learning techniques is that they need to be trained separately for each state point, thus requiring that training sets already exist at each time and temperature. Transferability to lower temperature has been analyzed in one of the first ML applications [27, 28] . It was shown that for GNNs some correlation between structure and dynamics persists when applying the trained networks to different temperatures","cbCaikVZVRHVvxa9","https://ap.wps.com/l/cbCaikVZVRHVvxa9","pdf",11035142,1,15,"English","en",105,"# Introduction\n## Glass transition and dynamic heterogeneity\n## Structural order parameters and ML approaches\n## Motivation and article aim\n# Transferable ML framework and applications\n## Training strategy and predictive range\n## Equilibrium structures near Tg and beyond-MD analysis\n## Organization of the manuscript","[{\"question\":\"What does the transferable machine learning model predict?\",\"answer\":\"It predicts structural relaxation and dynamic heterogeneity from amorphous supercooled liquid structures across broad temperature and time scales.\"},{\"question\":\"How is the experimental glass transition temperature Tg handled in this work?\",\"answer\":\"The model is transferred to predict relaxation down to Tg, a regime where structural relaxation cannot be analyzed using molecular dynamics simulations.\"},{\"question\":\"What key change is observed for dynamic heterogeneity near Tg?\",\"answer\":\"The strength, geometry, and characteristic length scale of dynamic heterogeneity evolve much more slowly near Tg than at higher temperatures.\"}]","Dynamic heterogeneity at the experimental glass transition predicted by transferable machine learning - Research overview | PDF",1785732321,38,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"dynamic-heterogeneity-at-the-experimental-glass-transition-predicted-by-transferable-machine-learning-research-overview","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/dynamic-heterogeneity-at-the-experimental-glass-transition-predicted-by-transferable-machine-learning-research-overview/120847/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What does the transferable machine learning model predict?","Question",{"text":75,"@type":76},"It predicts structural relaxation and dynamic heterogeneity from amorphous supercooled liquid structures across broad temperature and time scales.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is the experimental glass transition temperature Tg handled in this work?",{"text":80,"@type":76},"The model is transferred to predict relaxation down to Tg, a regime where structural relaxation cannot be analyzed using molecular dynamics simulations.",{"name":82,"@type":73,"acceptedAnswer":83},"What key change is observed for dynamic heterogeneity near Tg?",{"text":84,"@type":76},"The strength, geometry, and characteristic length scale of dynamic heterogeneity evolve much more slowly near Tg than at higher temperatures.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]