[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119039-en":3,"doc-seo-119039-105":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},119039,137441390410,"Hazel","https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984",8,"Research & Report","Using Machine Learning to Improve PDF Uncertainties","Parton distribution functions (PDFs) drive a major share of theoretical uncertainty in precision measurements at the Large Hadron Collider, including top-quark observables. This study targets the high momentum-fraction region by applying machine learning to tt+jet production, selecting events likely to originate from high-momentum initial gluons. A multilayer perceptron filters events using kinematic four-vectors, and ePump updates PDF uncertainty bands using the resulting differential pseudo-data. Filtering substantially reduces uncertainty in the high-x gluon region, with dramatic improvement in a best-case scenario under 1% systematics.","arXiv :2401 . 13050v1 [hep-ph] 23 Jan 2024  \nUsing Machine Learning to Improve PDF Uncertainties  \nJason P. Gombas, Reinhard Schwienhorst, Binbin Dong, and Jarrett  \nFein  \nDepartment of Physics and Astronomy  \nMichigan State University  \nParton Distribution Functions (PDFs) contribute significantly to the uncertainty on the determination of the top-quark pole mass and other precision measurements at the Large Hadron Collider (LHC) . It is crucial to understand these uncertainties and reduce them to obtain the next generation of precision measurements at the LHC. The region of high momentum fraction offers an opportunity to make improvements to the PDFs. This study uses machine learning techniques in tt production to target this region of the PDF set and has potential to significantly reduce its uncertainty.  \nPRESENTED AT  \n16th International Workshop on Top Quark Physics  \n(Top2023), 24–29 September, 2023  \n1 Introduction  \nWith the upcoming high luminosity large hadron collider (HL-LHC) [1], the next generation of precision measurements will be made. These measurements will be extremely precise, and will require lower theoretical uncertainties. Parton distribution functions (PDFs) are becoming the more dominant theoretical uncertainty in measurements like top quark pair production. However, many other measurements will also require reduced PDF uncertainties [2] .  \nColliders that will offer useful data that can significantly reduce PDF uncertainty, like the Electron Ion Collider [3], are far in the future. In the meantime, PDF uncertainty can be reduced using HL-LHC data.  \nMachine learning techniques can be used to pre-process the data to distill useful information to reduce specific regions of the PDFs with high uncertainty. Previous techniques involved using one, two or three dimensions of the high dimensional phase space of collider data to add to the global PDF fit [4] . Variables like rapidity and pZ of the top are typical choices. These variables do not include all the information. This proceeding will show current progress and some of preliminary results.  \n2 Simulation and Pseudo-Data  \nA sample of tt plus one jet events with a center of mass energy of 14 TeV was generated using Madgraph at next-to-leading order (NLO) [5] . A total of 7.5 million events were generated. The PDF set that was selected to study in detail was the CT18NLO PDF set [6] . The study looks at the truth level of ttj events without decaying the top quarks. The aim of this study is to constrain the high x region of the gluon PDF. ttj has been shown to be a process that has good potential to reduce this region of the PDFs [7] .  \n3 A Machine Learning Technique  \nTo test the idea of using machine learning to improve PDF fits, a MLP was developed to separate events with an  initial gluon parton that had greater than 2 TeV longitudinal momentum. These ttj events were considered signal. Events with less than 2 TeV longitudinal momentum were considered background. The inputs to the MLP were the kinematic 4-vectors of the final state particles (ttj) . Decent separation was achieved which, not surprisingly, indicates that there is information about the initial colliding partons (flavor and initial momentum) in just the kinematics of the final state particles. If the MLP output score was higher than 0 .7, it was considered signal and passed the MLP ”filter”. The MLP output scores can be seen in Fig. 1.  \nFigure 1: The MLP output scores for the trained MLP. Events that are closer to 1 are events that the MLP predicts to have an initial gluon parton whose initial momentum is greater than 2 TeV. Events that are closer to 0 are any other event. The inputs to this MLP are the kinematic 4-vectors of the final state ttj. There are 3 fully connected hidden layers. The peak at about 0.6 is currently not well understood.  \n4 PDF Update  \nTwo different differential distributions were created to compare how different methods can reduce PDF uncertainty. One histogram","cbCaic9fnxh0jxdC","https://ap.wps.com/l/cbCaic9fnxh0jxdC","pdf",701079,1,6,"English","en",105,"# Introduction\n## Simulation and Pseudo-Data\n## A Machine Learning Technique\n## PDF Update\n## Results\n## Outlook and Conclusions","[{\"question\":\"Why are PDF uncertainties important for LHC precision measurements?\",\"answer\":\"PDF uncertainties dominate theoretical uncertainty in several measurements such as top-quark pair production, making it crucial to understand and reduce them for next-generation precision at the HL-LHC.\"},{\"question\":\"How does the study use machine learning to improve PDF fits?\",\"answer\":\"A multilayer perceptron is trained to distinguish tt+jet events initiated by a gluon with longitudinal momentum greater than 2 TeV from events below that threshold, using kinematic four-vectors as inputs.\"},{\"question\":\"What approach is used to update PDF uncertainty bands and compare methods?\",\"answer\":\"Two sets of differential pseudo-data distributions are created—one using only events passing the MLP filter and one using all events—and these are fed into ePump to quantify how each constrains PDF uncertainty bands.\"}]","Using Machine Learning to Improve PDF Uncertainties | PDF",1785722038,15,{"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},"using-machine-learning-to-improve-pdf-uncertainties","",{"@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/using-machine-learning-to-improve-pdf-uncertainties/119039/",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},"Why are PDF uncertainties important for LHC precision measurements?","Question",{"text":75,"@type":76},"PDF uncertainties dominate theoretical uncertainty in several measurements such as top-quark pair production, making it crucial to understand and reduce them for next-generation precision at the HL-LHC.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the study use machine learning to improve PDF fits?",{"text":80,"@type":76},"A multilayer perceptron is trained to distinguish tt+jet events initiated by a gluon with longitudinal momentum greater than 2 TeV from events below that threshold, using kinematic four-vectors as inputs.",{"name":82,"@type":73,"acceptedAnswer":83},"What approach is used to update PDF uncertainty bands and compare methods?",{"text":84,"@type":76},"Two sets of differential pseudo-data distributions are created—one using only events passing the MLP filter and one using all events—and these are fed into ePump to quantify how each constrains PDF uncertainty bands.","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},[93,97,101,105,110,114,119,122,127,130,134],{"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":21,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"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"]