[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120984-en":3,"doc-seo-120984-105":29,"detail-sidebar-cat-0-en-105":90},{"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":11,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":26,"seo_description":14,"update_tm":27,"read_time":28},120984,137441390410,"Hazel","https://ap-avatar.wpscdn.com/avatar/2000252f4ab5702993?_k=1776741390130283984",8,"Research & Report","Simulating Hadronization with Machine Learning - Overview","Hadronization is a crucial, non-perturbative step in Monte Carlo event generation, turning quarks and gluons into observable hadrons and strongly affecting comparison with experimental data. The work presents MLhad, a machine-learning approach that learns latent-space representations trained to match a user-defined target distribution via sliced-Wasserstein distance, then decodes them to physical observables. Results reproduce pion multiplicities and kinematic distributions from Pythia 8, while an updated normalizing-flow architecture supports non-pion hadrons and reweighting to reduce computing time.","Simulating Hadronization with Machine Learning  \nMichael K. Wilkinson 1 , ∗ on behalf of the MLhad collaboration.  \n1Department of Physics, University of Cincinnati, Cincinnati, Ohio 45221, USA  \nAbstract. Hadronization is an important part of physics modeling in Monte Carlo event generators, where quarks and gluons are bound into physically observable hadrons. Today’s generators rely on finely-tuned phenomenological models, such as the Lund string model; while these models have been quite successful overall, there remain phenomenological areas where they do not match data well. A machine-learning-based alternative called MLhad, intended ultimately to be data-trainable, can simulate hadronization by encoding latentspace vectors, trained to be distributed according to a user-defined distribution using the sliced-Wasserstein distance in the loss function, then decoding them.  \nThe multiplicities and cumulative kinematic distributions of pions generated  \nwith MLhad in this way match those generated using Pythia 8.  \nWhile this architecture has been successful, an alternative using normalizing flows is convenient for generating non-pion hadrons and for taking advantage of reweighting techniques to reduce computing time. Combined with new methods for reweighting the output of phenomenological models, this updated architecture should prove convenient for comparing the output of MLhad and of empirical models.  \n1 Introduction  \nThe simulation of particle collisions can be considered in three blocks: (1) the hard process,(2) the parton shower, and (3) hadronization, as shown in figure 1 . The hard process is the initial high-energy interaction between partons (quarks, gluons, electrons, etc.); hadronization is the combination of quarks and gluons into hadrons; and the parton shower, also known as \"evolution\", corrects the hard process with additional quark and gluon emissions. The hard process and parton shower are well-determined by Quantum Chromo-Dynamics (QCD) through perturbative calculations, but hadronization is in a non-perturbative regime and must be simulated through the use of phenomenological models [1] .  \nThere are two widely used models used to simulate hadronization in general-purpose event generators: (1) the Lund string model [2] and (2) the cluster model [3], both illustrated in figure 1 . In the string model (used by Pythia 8 [4]), partons are connected via QCD color strings with linear potential; these strings are then iteratively split, emitting hadrons as long as there is enough energy to do so. In the cluster model (used by Herwig 7 [5, 6]), partons are pre-confined into proto-clusters, which then split via two-body decay. These splittings are not determined by QCD (in either model), presenting an opportunity for contributions from Machine Learning (ML) .  \n∗[e-mail: michael.wilkinson@uc.edu](e-mail: michael.wilkinson@uc.edu)  \n© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ([https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)).  \nFigure 1. (Left) the three blocks of event generation; the abbreviations refer to Multi-Parton Interaction (MPI), Final-State Radiation (FSR), Initial-State Radiation (ISR), and Quantum Electro-Dynamics (QED); the cross-section of the hard process is dˆσ0. (Right) the two primary hadronization models (used in block 3); in step 1, a quark qi and an anti-quark ¯qi , each with energy E and three-momentum p, are connected either by a linear QCD potential (string model) or by pre-confinement (cluster model); instep 2, this connection is iteratively broken by the creation of a new quark qj and anti-quark ¯qj , which combine with qi and ¯qi to form one or more hadrons with energies Eh and three-momenta ph.  \nMLhad is a novel ML-based approach to simulating hadronization using Neural Networks (NNs) . It is currently trained on Pythia 8 simulations and r","cbCairIAOjKtVwrN","https://ap.wps.com/l/cbCairIAOjKtVwrN","pdf",3767904,1,"English","en",105,"# Introduction\n# MLhad with cSWAE\n## cSWAE Architecture","[{\"question\":\"Why is hadronization important in event generators?\",\"answer\":\"Hadronization converts quarks and gluons into physically observable hadrons, and it lies in a non-perturbative regime that must be modeled in Monte Carlo generators.\"},{\"question\":\"How does MLhad simulate hadronization?\",\"answer\":\"MLhad encodes observables into latent-space vectors, trains them so their distribution matches a target using sliced-Wasserstein distance, and then decodes them back to simulated hadron observables.\"},{\"question\":\"What improvements are described beyond the original MLhad architecture?\",\"answer\":\"An updated architecture using normalizing flows is described to conveniently generate non-pion hadrons and to exploit reweighting techniques for reducing computing time.\"}]","Simulating Hadronization with Machine Learning - Overview | PDF",1785733173,20,{"code":4,"msg":30,"data":31},"ok",{"site_id":23,"language":22,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":85,"head_meta":87,"extra_data":89,"updated_unix":27},"simulating-hadronization-with-machine-learning-overview","",{"@graph":35,"@context":84},[36,53,67],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/simulating-hadronization-with-machine-learning-overview/120984/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":22,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"Why is hadronization important in event generators?","Question",{"text":74,"@type":75},"Hadronization converts quarks and gluons into physically observable hadrons, and it lies in a non-perturbative regime that must be modeled in Monte Carlo generators.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does MLhad simulate hadronization?",{"text":79,"@type":75},"MLhad encodes observables into latent-space vectors, trains them so their distribution matches a target using sliced-Wasserstein distance, and then decodes them back to simulated hadron observables.",{"name":81,"@type":72,"acceptedAnswer":82},"What improvements are described beyond the original MLhad architecture?",{"text":83,"@type":75},"An updated architecture using normalizing flows is described to conveniently generate non-pion hadrons and to exploit reweighting techniques for reducing computing time.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,119,122,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":28,"slug":125},9,"Religion & Spirituality","religion-spirituality",{"id":28,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":28,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]