[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125551-en":3,"doc-seo-125551-105":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},125551,1099513958762,"Logic","https://ap-avatar.wpscdn.com/avatar/1000023916a998db790?x-image-process=image/resize,m_fixed,w_180,h_180&k=1784791008015729253",8,"Research & Report","Evaluating Approaches for On-the-fly Machine Learning Interatomic Potential for Activated Mechanisms","Universal machine-learning interatomic potentials face rapidly increasing cost when materials become alloys and disordered or heterogeneous systems, because reliable description requires covering a growing set of local environments. This study evaluates benefits of specific versus general potentials for activated mechanisms in solid-state materials, testing three machine-learning fitting approaches using a Moment Tensor Potential within ARTn sampling. Targeted on-the-fly integration with ARTn yields the highest precision for energetic and geometric activated barriers while staying cost-effective, expanding solvable high-accuracy ML problems.","arXiv :2301 .08630v1 [ cond-mat .mtrl-sci ] 20 Jan 2023  \nEvaluating approaches for on-the-fly machine learning interatomic potential for activated mechanisms sampling with the activation-relaxation technique nouveau  \nEugène Sanscartier, 1 Félix Saint-Denis, 1 Karl-Étienne Bolduc, 1 and Normand Mousseau 1  \n1 Département de physique and Regroupement québécois sur les matériaux de pointe, Université de Montréal, Case Postale 6128, Succursale Centre-ville, Montréal, Québec H3C 3J7, Canada  \n(Dated: January 23, 2023)  \nIn the last few years, much efforts have gone into developing universal machine-learning potentials able to describe interactions for a wide range of structures and phases. Yet, as attention turns to more complex materials including alloys, disordered and heterogeneous systems, the challenge of providing reliable description for all possible environment become ever more costly. In this work, we evaluate the benefits of using specific versus general potentials for the study of activated mechanisms in solid-state materials. More specifically, we tests three machine-learning fitting approaches using the moment-tensor potential to reproduce a reference potential when exploring the energy landscape around a vacancy in Stillinger-Weber silicon crystal and silicon-germanium zincblende structure using the activation-relaxation technique nouveau (ARTn) . We find that a a targeted on-the-fly approach specific and integrated to ARTn generates the highest precision on the energetic and geometry of activated barriers, while remaining cost-effective. This approach expands the type of problems that can be addressed with high-accuracy ML potentials.  \nI. INTRODUCTION  \nAs computational materials scientists turn to attention to ever more complex systems, they are faced with two major challenges : (i) how to describe correctly their physics and (ii) how to reach the appropriate size and time scale to capture the properties of interest. The first challenge is generally solved by turning to ab initio methods, 1 that allow the solution Heisenberg’s equation with reasonably controlled approximations. Theses approaches, however, suffer from N4 scaling which limits their application to small system sizes and short timescales. The second challenge is met by a variety of methods that cover different scales. Molecular dynamics2 , for example, which directly solves Newton’s equation, accesses typical time scales between picoseconds and microseconds, at the very best. Other approaches, such as lattice3,4 and off-lattices kinetic Monte-Carlo5,6 , by focusing on physically relevant mechanisms, can extend this time scale to seconds and more, as long the diffusion takes place through activated processes. Even though these methods are efficient, each trajectory can require hundreds of thousands to millions of forces evaluations, which becomes too costly with ab initio approaches, forcing modellers to use empirical potentials in spite of their incapacity at describing correctly complex environments.  \nBuilding on ab initio energy and forces, machinelearned potentials7–10 open the door to lifting some of this difficulties, by offering much more reliable physics asa small fraction of the cost of ab initio evaluations.  \nSince their introduction, ML potentials have been largely coupled with MD and focusing on the search for universal potentials, able to describe a full range of structures and phases for a given material 11–13 . As we turn to more complex systems such as alloys and disordered and heterogeneous systems, it becomes more and more difficult to generate such universal potentials, since the  \nnumber of possible environments grows rapidly with this complexity. In this context, the development of specific potentials, with on-the-fly learning that makes it possible to adapt to new environments, becomes a strategy worth exploring.  \nIn this work, we focus on the construction of machinelearned potentials adapted to the sampling of energy landscape dominated by activated","cbCaigCRUEdvIi6U","https://ap.wps.com/l/cbCaigCRUEdvIi6U","pdf",657729,1,12,"English","en",105,"# Introduction\n## Challenges in modeling complex materials\n## Machine-learned potentials and universality limits\n## Activated mechanisms and ART/ARTn sampling\n# Methodology\n## ML potential (Moment Tensor Potential)\n## Learning procedures compared","[{\"question\":\"Why is universality difficult for machine-learning interatomic potentials in complex materials?\",\"answer\":\"As alloys and disordered or heterogeneous systems become more complex, the number of possible local environments grows rapidly, making universal reliable coverage increasingly costly to achieve.\"},{\"question\":\"What is ARTn used for in this study?\",\"answer\":\"ARTn (activation-relaxation technique nouveau) samples the energy landscape around activated mechanisms by identifying local minima and first-order saddle points associated with diffusion.\"},{\"question\":\"Which fitting strategy provides the best precision for activated barriers?\",\"answer\":\"The targeted on-the-fly approach specifically integrated to ARTn delivers the highest precision in both energetic and geometric properties of activated barriers while remaining cost-effective.\"}]","Evaluating Approaches for On-the-fly Machine Learning Interatomic Potential for Activated Mechanisms | PDF",1785899819,30,{"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},"evaluating-approaches-for-on-the-fly-machine-learning-interatomic-potential-for-activated-mechanisms","",{"@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/evaluating-approaches-for-on-the-fly-machine-learning-interatomic-potential-for-activated-mechanisms/125551/",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","2026-08-05",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},"Why is universality difficult for machine-learning interatomic potentials in complex materials?","Question",{"text":75,"@type":76},"As alloys and disordered or heterogeneous systems become more complex, the number of possible local environments grows rapidly, making universal reliable coverage increasingly costly to achieve.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is ARTn used for in this study?",{"text":80,"@type":76},"ARTn (activation-relaxation technique nouveau) samples the energy landscape around activated mechanisms by identifying local minima and first-order saddle points associated with diffusion.",{"name":82,"@type":73,"acceptedAnswer":83},"Which fitting strategy provides the best precision for activated barriers?",{"text":84,"@type":76},"The targeted on-the-fly approach specifically integrated to ARTn delivers the highest precision in both energetic and geometric properties of activated barriers while remaining cost-effective.","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},[93,97,101,105,110,115,120,122,127,130,134],{"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":29,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]