[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117077-en":3,"doc-seo-117077-105":30,"detail-sidebar-cat-0-en-105":95},{"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},117077,4810365810221,"Aurora","https://ap-avatar.wpscdn.com/davatar_155a257f0dc6eb9ab79c44ca47cae57d",8,"Research & Report","Machine Learning Force Fields Under the Microscope - Stability, Reliability and Performance Analysis - Dissertation","Machine Learning Force Fields (MLFF) enable accurate yet feasible predictions that approximate computationally expensive quantum calculations for molecules and materials. As MLFF architectures have grown in complexity, evaluating model quality has become harder because prediction errors are highly heterogeneous across configurational space, directly affecting stability and reliability in applications such as molecular dynamics. This dissertation introduces FFAST, cross-platform software with a graphical interface for detailed, resolution-flexible performance diagnostics. It also develops an unsupervised clustering strategy to assess predictive accuracy in distinct configurational regions and guide iterative training for improved stability and reduced force errors.","PhD-FSTM-2023-109  \nThe Faculty of Science, Technology and Medicine  \nDISSERTATION  \nDefence held on 25/10/2023 in Luxembourg to obtain the degree of  \nDOCTEUR DE L’UNIVERSITÉ DU LUXEMBOURG  \nEN PHYSIQUE  \nby  \nGrégory Cordeiro Fonseca  \nBorn on 30 May 1993 in Luxembourg  \nMACHINE LEARNING FORCE FIELDS UNDER THE  \nMICROSCOPE: STABILITY, RELIABILITY AND PERFORMANCE ANALYSIS  \nDissertation defence committee Dr Alexandre Tkatchenko, dissertation supervisor Professor, Université du Luxembourg  \nDr Ludger Wirtz, Chairman Professor, Université du Luxembourg  \nDr Gábor Csányi, Vice Chairman Professor, University of Cambridge  \nDr Matthias Rupp  \nResearch Group Leader, Luxembourg Institute of Science and Technology  \nDr O. Anatole von Lilienfeld Professor, University of Toronto  \niii  \nUNIVERSITY OF LUXEMBOURG  \nAbstract  \nFaculty of Science, Technology and Medicine  \nDepartment of Physics and Materials Science  \nDoctor in Physics  \nMachine Learning Force Fields Under the Microscope: Stability, Reliability and  \nPerformance Analysis  \nby Cordeiro Fonseca Gregory  \nMachine Learning Force Fields (MLFF) are a crucial tool for bringing the accuracy of computationally expensive quantum mechanical calculations to practically feasible applications on molecules and materials. Over time, the sophistication of MLFF architectures has increased to match the complexity of systems of ever-growing sizes. This increase in complexity comes with a higher need for in-depth analytical tools to properly assess a Machine Learning (ML) model’s quality. Even the most advanced models that showcase remarkably low overall prediction errors demonstrate highly heterogeneous predictive capabilities across the Configurational Space (CS) of a single system. In practice, these can significantly impact the reliability of a model as a high prediction error on a few key geometries can easily lead to e.g. the destabilisation of a molecular dynamics simulation.  \nIn this work, we provide a cross-platform software package designed to give a detailed view into the performance and shortcomings of an MLFF model, complete with an easy-to-use graphical user interface. Entitled FFAST (Force Field Analysis Software and Tools), this actively developed software enables any user to gauge the quality of many state-of-the-art ML architectures and infer potential pitfalls in  \niv  \npractical applications. Analytical tools are provided at any desired level of resolution, from average error metrics over entire datasets to assessments of prediction accuracies on an atom-by-atom basis.  \nTo provide an optimal compromise between detailed analysis and simplicity, a novel approach is developed to determine a model’s predictive capabilities across different regions of CS. This is achieved by employing methods from unsupervised learning to create clusters of qualitatively different configurations and calculate their respective prediction accuracies separately. This provides much-needed context to otherwise general error metrics and captures insightful details of the model’s capacity to reproduce e.g. important out-of-equilibrium mechanisms rarely frequented in the reference dataset. Furthermore, inhomogeneous error curves across clusters also provide information on which regions are likely poorly represented in a model’s training set.  \nThe potential of the aforementioned methods as well as FFAST are showcasedon example datasets of stachyose and docosahexaenoic acid (DHA) as well as a handful of smaller organic molecules. After successfully proceeding through atypical FFAST workflow with two state-of-the-art ML models (Nequip and MACE), it was quickly determined that carbons and oxygens near glycosidic bonds of the stachyose molecule have increased prediction errors. Furthermore, prediction errors on DHA rise as the molecule folds, with notably low accuracy on the carboxylic group at the edge of the molecule. Finally, the cluster prediction errors of a handful of small organic molecules were genera","cbCaibbOUpdcMN9B","https://ap.wps.com/l/cbCaibbOUpdcMN9B","pdf",13568325,1,155,"English","en",105,"# Abstract\n## FFAST software and diagnostic tools\n## Unsupervised clustering across configurational space\n## Case studies and iterative training for stability\n## Future research directions","[{\"question\":\"Why is heterogeneity in MLFF predictions important?\",\"answer\":\"Even with low average prediction errors, models can behave very differently across configurational space. Errors concentrated in a few geometries can reduce reliability and destabilize molecular dynamics simulations.\"},{\"question\":\"What is FFAST and what does it enable?\",\"answer\":\"FFAST is a cross-platform, actively developed software package with a graphical user interface. It provides analytical tools to inspect MLFF performance and shortcomings from dataset-level metrics down to atom-by-atom accuracy.\"},{\"question\":\"How does the dissertation assess predictive capability across different regions of configurational space?\",\"answer\":\"It develops an unsupervised-learning approach that clusters qualitatively different configurations and computes prediction accuracies separately for each cluster. This adds context missing from global error metrics and highlights poorly represented regions.\"},{\"question\":\"What improvements were observed after iterative training based on cluster underrepresentation?\",\"answer\":\"Models trained with this iterative, cluster-guided process show enhanced stability in molecular dynamics. They achieve force prediction improvements, including up to a two-fold reduction in root mean squared errors in problematic configurational regions.\"}]","Machine Learning Force Fields Under the Microscope - Stability, Reliability and Performance Analysis - Dissertation | PDF",1785673580,391,{"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":90,"head_meta":92,"extra_data":94,"updated_unix":28},"machine-learning-force-fields-under-the-microscope-stability-reliability-and-performance-analysis-dissertation","",{"@graph":36,"@context":89},[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/machine-learning-force-fields-under-the-microscope-stability-reliability-and-performance-analysis-dissertation/117077/",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","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81,85],{"name":72,"@type":73,"acceptedAnswer":74},"Why is heterogeneity in MLFF predictions important?","Question",{"text":75,"@type":76},"Even with low average prediction errors, models can behave very differently across configurational space. Errors concentrated in a few geometries can reduce reliability and destabilize molecular dynamics simulations.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is FFAST and what does it enable?",{"text":80,"@type":76},"FFAST is a cross-platform, actively developed software package with a graphical user interface. It provides analytical tools to inspect MLFF performance and shortcomings from dataset-level metrics down to atom-by-atom accuracy.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the dissertation assess predictive capability across different regions of configurational space?",{"text":84,"@type":76},"It develops an unsupervised-learning approach that clusters qualitatively different configurations and computes prediction accuracies separately for each cluster. This adds context missing from global error metrics and highlights poorly represented regions.",{"name":86,"@type":73,"acceptedAnswer":87},"What improvements were observed after iterative training based on cluster underrepresentation?",{"text":88,"@type":76},"Models trained with this iterative, cluster-guided process show enhanced stability in molecular dynamics. They achieve force prediction improvements, including up to a two-fold reduction in root mean squared errors in problematic configurational regions.","https://schema.org",{"og:url":52,"og:type":91,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":93,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":96},[97,101,105,109,114,119,124,127,132,135,139],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":106,"show_sort_weight":107,"slug":108},"Exam",70,"exam",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},5,"Comic",60,"comic",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},6,"Technology",50,"technology",{"id":120,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":122,"slug":123},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":125,"slug":126},30,"research-report",{"id":128,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":130,"slug":131},9,"Religion & Spirituality",20,"religion-spirituality",{"id":130,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":130,"slug":134},"World Cup","world-cup",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":136,"slug":138},10,"Lifestyle","lifestyle",{"id":140,"doc_module":4,"doc_module_name":46,"category_name":141,"show_sort_weight":110,"slug":142},19,"General","general"]