[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-124344-en":3,"doc-seo-124344-105":30,"detail-sidebar-cat-0-en-105":96},{"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},124344,962075114101,"Seraphina","https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165",8,"Research & Report","Generative Machine Learning for Simulation-based Inference in High Energy Physics - Dissertation","With the upcoming High-Luminosity LHC, collider data will grow dramatically, demanding major upgrades to existing simulation and inference workflows to avoid computational bottlenecks. This dissertation studies how generative machine learning can improve precision physics pipelines. It evaluates diffusion models and autoregressive transformers for fast, accurate LHC event generation, demonstrating surrogate simulation potential at percent-level precision. It then advances machine learning for the matrix element method by encoding transfer probabilities while keeping phase-space integration tractable. Finally, it explores high-dimensional, unbinned unfolding using generative models, benchmarks multiple methods on the same datasets, and introduces a transformer-enhanced diffusion model achieving state-of-the-art precision.","Generative Machine Learning for Simulation-based Inference in High Energy Physics  \nDissertation  \nNathan Hütsch  \nDissertation  \nsubmitted to the  \nCombined Faculty of Mathematics, Engineering and Natural Sciences of Heidelberg University, Germany  \nfor the degree of  \nDoctor of Natural Sciences  \nPut forward by  \nNathan Hütsch  \nborn in Frankfurt am Main, Germany  \nOral examination: 04.07.2025  \nGenerative Machine Learning for Simulation-based Inference in High Energy Physics  \nReferees: Dr. Anja Butter  \nProf. Dr. Ullrich Köthe  \nAbstract  \nWith the upcoming High-Luminosity LHC the volume of collider data will increase dramatically, leading to a new era of precision measurements. However, this also creates computational and methodological challenges. Established simulation and inference pipelines require significant upgrades to prevent them from becoming bottlenecks. This thesis investigates how generative machine learning can address these challenges. First, we investigate modern generative architectures, diffusion models and autoregressive transformers, for fast and accurate LHC event generation. We find that they can learn complex phase space distributions to percent-level precision, demonstrating their potential as surrogate simulators. Second, we advance the use of machine learning for the matrix element method, showing how generative networks can be used to encode the transfer probability and keep the phase space integration tractable. Finally, we explore high-dimensional, unbinned unfolding using generative models. We benchmark the performance of a range of methods on the same datasets and contribute several methodological advancements, including a transformer-enhanced diffusion model that achieves state-of-the-art precision.  \nZusammenfassung  \nMit dem bevorstehenden High-Luminosity LHC (HL-LHC) wird das Volumen der Kollisionsdaten drastisch zunehmen, was eine neue Ära von Präzisionsmessungen einläutet. Dies bringt jedoch auch erhebliche ressourcentechnische und methodische Herausforderungen mit sich. Die etablierten Simulations-und Inferenzstrategien müssen grundlegend weiterentwickelt werden, um nicht zum Engpass zu werden. In dieser Arbeit wird untersucht, wie generative Methoden des maschinellen Lernens zur Bewältigung dieser Herausforderungen beitragen können. Zunächst evaluieren wir moderne generative Architekturen, Diffusionsmodelle und autoregressive Transformer, im Hinblick auf eine schnelle und präzise Ereignissimulation am LHC. Wir zeigen, dass diese Modelle komplexe Phasenraumverteilungen mit einer Genauigkeit auf Prozentniveau erlernen können und damit als Ersatz-Simulatoren großes Potenzial besitzen. Im zweiten Teil erweitern wirden Einsatz von maschinellem Lernen für die Matrixelementmethode, indem wir zeigen, wie generative Netzwerke sowohl die Transferwahrscheinlichkeit modellieren als auch die Integration über den Phasenraum effizient gestalten können. Zuletzt widmen wir uns demungebinnten, hoch-dimensionalen Unfolding mittels generativer Modelle. Wir vergleichen die Leistung verschiedener Methoden auf denselben Datensätzen und leisten mehrere methodische Beiträge, darunter ein Diffusionsmodell mit Transformer-Architektur, das state-of-the-art Präzision erreicht.  \nContents  \nPreface iii  \n1 Introduction 1  \n2 LHC physics 5  \n2.1 Collider physics ................................. 7  \n2.1. 1 Cross sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7  \n2.1.2 Parton distribution functions . . . . . . . . . . . . . . . . . . . . . 8  \n2.2 The LHC simulation chain . . . . . . . . . . . . . . . . . . . . . . . . . . . 9  \n2.2.1 Hard-scattering event generation . . . . . . . . . . . . . . . . . . . 10  \n2.2.2 Parton showering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11  \n2.2.3 Hadronization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12  \n2.2.4 Detector effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12  \n2.3 Inference in LHC physics . . .","cbCaie68f2QyEyml","https://ap.wps.com/l/cbCaie68f2QyEyml","pdf",8700384,1,180,"English","en",105,"# Introduction\n# LHC physics\n## The LHC simulation chain\n## Inference in LHC physics\n# Machine Learning\n## Learning tasks\n## Simulation-based inference\n# Precision event generation with diffusion models and transformers\n## Novel generative networks\n# Precision-Machine Learning for the Matrix Element Method","[{\"question\":\"Why does the High-Luminosity LHC require changes to simulation and inference pipelines?\",\"answer\":\"The High-Luminosity LHC increases collider data volume dramatically, which creates computational and methodological challenges. Established pipelines must be upgraded to avoid becoming bottlenecks.\"},{\"question\":\"How do generative models improve LHC event generation?\",\"answer\":\"The dissertation evaluates diffusion models and autoregressive transformers to generate LHC events fast and accurately. The models learn complex phase-space distributions with percent-level precision, enabling surrogate simulation.\"},{\"question\":\"What role do generative networks play in the matrix element method?\",\"answer\":\"Generative networks are used to encode the transfer probability while keeping phase-space integration tractable. This extends machine learning applications within the matrix element method.\"},{\"question\":\"How is unfolding performed using generative models in this thesis?\",\"answer\":\"The work explores high-dimensional, unbinned unfolding with generative models. It benchmarks multiple approaches on the same datasets and reports methodological advancements, including a transformer-enhanced diffusion model.\"}]","Generative Machine Learning for Simulation-based Inference in High Energy Physics - Dissertation | PDF",1785821723,454,{"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":91,"head_meta":93,"extra_data":95,"updated_unix":28},"generative-machine-learning-for-simulation-based-inference-in-high-energy-physics-dissertation","",{"@graph":36,"@context":90},[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/generative-machine-learning-for-simulation-based-inference-in-high-energy-physics-dissertation/124344/",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","2026-08-05","2026-08-04",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82,86],{"name":73,"@type":74,"acceptedAnswer":75},"Why does the High-Luminosity LHC require changes to simulation and inference pipelines?","Question",{"text":76,"@type":77},"The High-Luminosity LHC increases collider data volume dramatically, which creates computational and methodological challenges. Established pipelines must be upgraded to avoid becoming bottlenecks.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How do generative models improve LHC event generation?",{"text":81,"@type":77},"The dissertation evaluates diffusion models and autoregressive transformers to generate LHC events fast and accurately. The models learn complex phase-space distributions with percent-level precision, enabling surrogate simulation.",{"name":83,"@type":74,"acceptedAnswer":84},"What role do generative networks play in the matrix element method?",{"text":85,"@type":77},"Generative networks are used to encode the transfer probability while keeping phase-space integration tractable. This extends machine learning applications within the matrix element method.",{"name":87,"@type":74,"acceptedAnswer":88},"How is unfolding performed using generative models in this thesis?",{"text":89,"@type":77},"The work explores high-dimensional, unbinned unfolding with generative models. It benchmarks multiple approaches on the same datasets and reports methodological advancements, including a transformer-enhanced diffusion model.","https://schema.org",{"og:url":52,"og:type":92,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":94,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":97},[98,102,106,110,115,120,125,128,133,136,140],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},"Exam",70,"exam",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},5,"Comic",60,"comic",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},6,"Technology",50,"technology",{"id":121,"doc_module":4,"doc_module_name":46,"category_name":122,"show_sort_weight":123,"slug":124},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":126,"slug":127},30,"research-report",{"id":129,"doc_module":4,"doc_module_name":46,"category_name":130,"show_sort_weight":131,"slug":132},9,"Religion & Spirituality",20,"religion-spirituality",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":134,"show_sort_weight":131,"slug":135},"World Cup","world-cup",{"id":137,"doc_module":4,"doc_module_name":46,"category_name":138,"show_sort_weight":137,"slug":139},10,"Lifestyle","lifestyle",{"id":141,"doc_module":4,"doc_module_name":46,"category_name":142,"show_sort_weight":111,"slug":143},19,"General","general"]