[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119202-en":3,"doc-seo-119202-105":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},119202,1099514068035,"Ezra","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",6,"Technology","Machine Learning Potentials - A Roadmap Toward Next-Generation Biomolecular Simulations","Machine learning potentials provide a unifying framework for molecular simulations across scales, aiming to enhance both accuracy and computational scalability from quantum chemistry to coarse-grained modeling. The discussion outlines how neural networks can learn complex correlations in high-dimensional spaces, improving efficiency while bridging atomistic detail and macroscopic behavior. Key limitations of classical force fields and ad hoc coarse-grained forms are contrasted with the promise of transferable neural network potentials. Challenges and future research directions in chemical biology and related applications are also emphasized.","MACHINE LEARNING POTENTIALS: A ROADMAP TOWARD NEXT-GENERATION BIOMOLECULAR SIMULATIONS  \nGianni De Fabrit iis1,2,3  \n1 Computational Science Laboratory, Universitat Pompeu Fabra, Barcelona Biomedical Research Park ( PRBB), C Dr. A iguader 88, 08003 Barcelona, Spain.  \n2 Acellera Therapeutics, 38350 Fremont Blvd 203, Fremont CA, 94536 USA.  \n3 ICREA, Passeig L luis Companys 23, 08010 Barcelona, Spain.  \n[Email: g.defabritiis@gmail.com](Email: g.defabritiis@gmail.com)  \nMachine learning potentials offer a revolutionary, unifyingframework for molecular simulations across scales, from quantum chemistry to coarse-grained models. Here, I explore their potential to dramatically improve accuracy and scalability in simulating complex molecular systems. I discuss key challenges that must be addressed to fully realize their transformative potential in chemical biology and related fields.  \nMachine learning ( ML) potentials are poised to revolutionize molecular simulations across multiple scales, leveraging the innate ability of neural networks to capture complex correlations in high-dimensional spaces. Here, I discuss how these advanced models can dramatically improve the accuracy and efficiency of simulations, from quantum-mechanical calculations to coarse-grained dynamics. By bridging the gap between atomistic detail and macroscopic behavior, ML potentials promise to unlock new insights into molecular processes, drug discovery, and materials design [ Duignan24] . I highlight recent successes, current challenges, and future directions in this rapidly evolving field, emphasizing the transformative potential of ML-driven simulations in chemical biology and related disciplines.  \nClassical atomistic molecular mechanics potentials are mathematical models used to describe the energy and forces between atoms in a molecular system. Unlike quantum chemistry methods, which explicitly treat electronic structures ( Figure 1a), these potentials  \ncoarse-grain the effects of electrons into simplified interaction terms between atomic centers ( Figure 1b) . Typically, they consist of bonded terms (describing bond stretching, angle bending, and torsional rotations) and non-bonded terms (such as van der Waals and electrostatic interactions) . This simplification allows for the simulation of much larger systems and longer timescales than quantum chemistry methods, but at the cost of reduced accuracy and the inability to model electronic processes. The parameters for these potentials are usually derived from experimental data or higher-level quantum calculations. While classical potentials have been invaluable in many areas of molecular modeling [ Lee09], their fixed functional forms limit their ability to capture complex, many-body interactions accurately across diverse chemical environments.  \nTo simulate even larger macromolecules and their assemblies researchers have developed different coarse-grained functional forms and parameterizations that further reduce computational complexity. Traditional coarse-grained force fields, group multiple heavy atoms into single interaction sites or \"beads\" using predefi ned [ Marrink07] and functional forms. These models significantly extend the accessible time and length scales of simulations, enabling the study of complex biological processes like protein folding, membrane dynamics, and macromolecular assembly. However, the ad hoc nature of their functional form and parameterization can limit transferability and accuracy.  \nNeural network potentials offer an exciting, unifying language for molecular simulations across scales. This paradigm shift allows us to view classical molecular mechanics and coarse-grained simulations as instances of t he same fundamental learning process, starting from quantum mechanical principles and progressively abstracting to coarser models ( Figure 1abc) . By framing potential energy surfaces as learnable functions, we can systematically derive models at various levels of granularity while maintaini","cbCaitUKB3Nn7LPX","https://ap.wps.com/l/cbCaitUKB3Nn7LPX","pdf",1912588,1,14,"English","en",105,"# Introduction\n## Why machine learning potentials matter across scales\n## Classical potentials and their limitations\n## Coarse-grained models and transferability\n## Neural network potentials as a unifying framework\n## From quantum structure to multi-scale dynamics","[{\"question\":\"What are machine learning potentials and what problem do they target?\",\"answer\":\"Machine learning potentials are neural-network-based models that approximate potential energy surfaces to enable molecular simulations with improved accuracy and scalability across multiple resolution levels.\"},{\"question\":\"How do classical atomistic and coarse-grained potentials differ from machine learning approaches?\",\"answer\":\"Classical atomistic potentials use fixed functional forms that simplify electronic effects, while coarse-grained models reduce complexity further using predefined parameterizations; both can limit accuracy and transferability. Machine learning potentials aim for more transferable, learned interactions across environments.\"},{\"question\":\"How can neural network potentials bridge different molecular simulation scales?\",\"answer\":\"By framing potential energy surfaces as learnable functions, neural networks can be trained at multiple granularities, supporting transitions from electronic structure calculations to atomistic and then mesoscale/macroscale models while preserving quantitative agreement.\"}]","Machine Learning Potentials - A Roadmap Toward Next-Generation Biomolecular Simulations | PDF",1785723079,35,{"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},"machine-learning-potentials-a-roadmap-toward-next-generation-biomolecular-simulations","",{"@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/technology/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-potentials-a-roadmap-toward-next-generation-biomolecular-simulations/119202/",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 are machine learning potentials and what problem do they target?","Question",{"text":75,"@type":76},"Machine learning potentials are neural-network-based models that approximate potential energy surfaces to enable molecular simulations with improved accuracy and scalability across multiple resolution levels.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How do classical atomistic and coarse-grained potentials differ from machine learning approaches?",{"text":80,"@type":76},"Classical atomistic potentials use fixed functional forms that simplify electronic effects, while coarse-grained models reduce complexity further using predefined parameterizations; both can limit accuracy and transferability. Machine learning potentials aim for more transferable, learned interactions across environments.",{"name":82,"@type":73,"acceptedAnswer":83},"How can neural network potentials bridge different molecular simulation scales?",{"text":84,"@type":76},"By framing potential energy surfaces as learnable functions, neural networks can be trained at multiple granularities, supporting transitions from electronic structure calculations to atomistic and then mesoscale/macroscale models while preserving quantitative agreement.","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},[93,97,101,105,110,113,118,123,128,131,135],{"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":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":111,"slug":112},50,"technology",{"id":114,"doc_module":4,"doc_module_name":46,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":119,"doc_module":4,"doc_module_name":46,"category_name":120,"show_sort_weight":121,"slug":122},8,"Research & Report",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"]