[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118287-en":3,"doc-seo-118287-105":30,"detail-sidebar-cat-0-en-105":92},{"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":20,"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},118287,7971461740886,"Theodore","https://ap-avatar.wpscdn.com/davatar_3d24733baf745e90a7e4bdd5f77d97b2",8,"Research & Report","Systematic assessment of various universal machine-learning interatomic potentials","Machine-learning interatomic potentials have transformed materials modeling by enabling ab initio–quality simulations at far larger time and length scales. Universal machine-learning models further reduce friction by avoiding bespoke training and validation for each new material. This paper reviews and evaluates four universal ML interatomic potentials based on graph neural network architectures, focusing on transferability across chemical systems. The assessment uses data from a recent DFT verification study and from the Materials Project. Recommendations and selection guidance are provided, alongside discussion of improvements for current methodologies.","arXiv :2403 .05729v3 [ cond-mat .mtrl-sci ] 20 Jul 2024  \nSystematic assessment of various universal machine-learning interatomic potentials  \nHaochen Yu1  Matteo Giantomassi1  Giuliana Materzanini1  Junjie Wang2  Gian-Marco Rignanese1,2,3  \n1Institute of Condensed Matter and Nanosciences, Université catholique de Louvain, 1348  \nLouvain-la-Neuve, Belgium  \n2 State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an, Shaanxi 710072, Republic of China  \n3WEL Research Institute, 1300 Wavre, Belgium  \nCorrespondence  \nCorresponding author Gian-Marco Rignanese, Email: [gian-marco.rignanese@uclouvain.be](gian-marco.rignanese@uclouvain.be)  \nAbstract  \nMachine-learning interatomic potentials have revolutionized materials modeling at the atomic scale. Thanks to these, it is now indeed possible to perform simulations of ab initio quality over very large time and length scales. More recently, various universal machine-learning models have been proposed as an out-of-box approach avoiding the need to train and validate specific potentials for each particular material of interest. In this paper, we review and evaluate four different universal machine-learning interatomic potentials (uMLIPs), all based on graph neural network architectures which have demonstrated transferability from one chemical system to another. The evaluation procedure relies on data both from a recent verification study of densityfunctional-theory implementations and from the Materials Project. Through this comprehensive evaluation, we aim to provide guidance to materials scientists in selecting suitable models for their specific research problems, offer recommendations for model selection and optimization, and stimulate discussion on potential areas for improvement in current machine-learning methodologies in materials science.  \nK E Y W O R D S  \nuniversal machine-learning interatomic potentials, verification, machine learning, phonons, formation energy, geometry optimization  \n1  INTRODUCTION  \nMaterials simulations at the atomic scale are the backbone of computational materials design and discovery. They rely on the Born Oppenheimer approximation, in which the electrons follow the nuclear motion adiabatically, so that the potential governing the nuclei consists of the electronic energies as a function of the nuclear positions, called \"potential energy surface\" (PES) . Knowledge of the PES allows the identification of stable and metastable atomic configurations from minimum energy search, or the determination of materials properties as thermodynamical averages from molecular dynamics simulations 1. The utility of atom-based materials simulations is thus intimately related to the generation of accurate PESs, which has been possible in the last decades thanks to the advent of density-functional theory (DFT) 2,3,4,5 . Nonetheless, this ab initio approach relies on the quantum mechanical solution of the electronic problem whose computational cost scales cubically with system size and can therefore become unaffordable in  \nvarious significant cases of technological interest such as amorphous solids, interfaces, surfaces, etc. At the other end of the simulation approaches, parametrized approximations of the Born-Oppenheimer PES, known as empirical analytical potentials, or force fields, or \"classical” interatomic potentials, have been widely used especially for large-scale materials studies 6. Unfortunately, in particular when complex electron interactions are involved (as in chemical reactions or phase transitions) these approaches cannot usually achieve DFT accuracy, and in addition they have limited applicability and transferability. They cannot therefore be considered as a drop-in replacement for standard ab initio methods. In this context, machine-learning interatomic potentials (MLIPs) have emerged as an in-between solution with computational cost similar to the empirical analytical potentials, but with the promise of achieving","cbCaidPRw9NDPnzV","https://ap.wps.com/l/cbCaidPRw9NDPnzV","pdf",8035738,1,13,"English","en",105,"# Introduction\n## Background: PES and atomistic simulations\n## From empirical potentials to MLIPs\n## Universal MLIPs and graph neural network models\n# Abstract and study scope\n## Review and evaluation goals\n## Data sources and evaluation procedure\n# Keywords and target properties\n## Phonons, formation energy, and geometry optimization","[{\"question\":\"What problem do universal machine-learning interatomic potentials aim to solve?\",\"answer\":\"They aim to provide an out-of-box alternative that avoids training and validating separate interatomic potentials for each specific material.\"},{\"question\":\"How are the universal MLIPs evaluated in this work?\",\"answer\":\"The evaluation combines data from a recent verification study of density-functional-theory implementations and datasets from the Materials Project.\"},{\"question\":\"What types of models does the paper focus on?\",\"answer\":\"It focuses on four universal ML interatomic potentials built on graph neural network architectures, designed to demonstrate transferability across different chemical systems.\"}]","Systematic assessment of various universal machine-learning interatomic potentials | PDF",1785682810,33,{"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":87,"head_meta":89,"extra_data":91,"updated_unix":28},"systematic-assessment-of-various-universal-machine-learning-interatomic-potentials","",{"@graph":36,"@context":86},[37,54,69],{"@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/systematic-assessment-of-various-universal-machine-learning-interatomic-potentials/118287/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-06","2026-08-02",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What problem do universal machine-learning interatomic potentials aim to solve?","Question",{"text":76,"@type":77},"They aim to provide an out-of-box alternative that avoids training and validating separate interatomic potentials for each specific material.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How are the universal MLIPs evaluated in this work?",{"text":81,"@type":77},"The evaluation combines data from a recent verification study of density-functional-theory implementations and datasets from the Materials Project.",{"name":83,"@type":74,"acceptedAnswer":84},"What types of models does the paper focus on?",{"text":85,"@type":77},"It focuses on four universal ML interatomic potentials built on graph neural network architectures, designed to demonstrate transferability across different chemical systems.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":46,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":46,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":46,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":46,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":46,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]