[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119175-en":3,"doc-seo-119175-105":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},119175,8796095461610,"Oliver","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Particle Jet Classification Using Edge Machine Learning - Master’s Thesis 2024","This master’s thesis studies machine learning models for particle jet classification using Large Hadron Collider (LHC) data, with attention to computational constraints expected from the HL-LHC high-luminosity upgrade. Deep learning methods are developed to produce results that can be adapted for deployment on field-programmable gate arrays (FPGAs). The core approach uses a Deep Sets B-tagging system suited for unordered jet constituents as in the CMS experiment. Quantization and pruning are integrated during training to meet FPGA hardware limits, making subsequent deployment feasible.","PARTICLE JET CLASSIFICATION USING EDGE MACHINE LEARNING  \nLappeenranta-Lahti University of Technology LUT  \nMaster’s Program in Computational Engineering, Master’s Thesis 2024  \nSaqib Saghir  \nExaminers: Professor Lasse Lensu  \nDr. Henri Petrow  \nABSTRACT  \nLappeenranta-Lahti University of Technology LUT School of Engineering Science  \nComputational Engineering  \nSaqib Saghir  \nParticle jet classi􀀂cation using edge machine learning  \nMaster’s thesis 2024  \n60 pages, 25 􀀂gures, 11 tables  \nExaminers: Professor Lasse Lensu and Dr. Henri Petrow  \nKeywords: edge machine learning, particle jet classi􀀂cation, particle physics, deep neural network, model pruning, model quantization, 􀀂eld-programmable gate array, large hadron collider, compact muon solenoid, level-1 trigger.  \nThis thesis focuses on studying machine learning models for particle jet classi􀀂cation based on Large Hadron Collider (LHC) data. The analysis of the computational challenges of the high-luminosity upgrade (HL-LHC) was aimed at by means of deep learning, with the results being adaptable for the deployment on 􀀂eld-programmable gate arrays (FPGAs) HL-LHC is an upgrade to the present LHC that will boost its luminosity, needed for physicists to explore basic particles and forces with unprecedented granularity and possibly make groundbreaking discoveries.  \nThe used machine learning model for the particle jet classi􀀂cation is based on the Deep Sets B-tagging system. The use of Deep Sets is promising as it is capable of dealing with unordered inputs. Such is the case for the compact muon solenoid (CMS) experiment at LHC, where the order of jet constituents is not 􀀂xed and can vary. The model is quantized and pruned as part of the training so that 􀀂ts to the FPGA-based hardware constraints. Due to the ef􀀂ciency of the quantization and pruning methods, the task of subsequently deploying the model on the FPGA can be considered feasible.  \nACKNOWLEDGEMENTS  \nFirst of all special thanks to my supervisors, Dr. Henri Petrow and Professor Lasse Lensu for their immense support and mentorship throughout this research endeavor. Of course, without their effort and guidance, it would not be possible. Their constructive feedback and deep insights were really helpful. Their guidance in utilizing new tools and arranging the resources required for this thesis is remarkable.  \nI am extremely grateful to the faculty of the Department of Computational Engineering and personnel of the Lappeenranta-Lahti University of Technology LUT for their support and resources throughout my academic pursuits.  \nI also would like to acknowledge, especially my brother \"Aqib Saghir\" and my family for their continuous support and assistance. Their moral support throughout this postgraduate study, it has served as a source of power and inspiration.  \nDuring the preparation of this master’s thesis, Particle Jet Classi􀀂cation Using Edge Machine Learning, I acknowledge the use of \"Grammarly\" to correct grammar mistakes in the text for improvement. After using \"Grammarly\", I have reviewed and edited the content and take full responsibility for the thesis content.  \nLappeenranta, June 14, 2024  \nSaqib Saghir  \n4  \nLIST OF ABBREVIATIONS  \nAI Arti􀀂cial Intelligence  \nANN Arti􀀂cial Neural Networks  \nASIC Application Speci􀀂c Integrated Circuit  \nAUC Area Under the Curve  \nBDT Boosted Decision Trees  \nCMSSW Compact Muon Solenoid Software  \nCMS Compact Muon Solenoid CPU Central processing unit DNN Deep Neural Networks DS Deep Sets  \nDW-GNN Distance Weighted Graph Neural Network  \nEIC Electron Ion Collider  \nETS Event Samples  \nFEB Front End Board  \nFPGA Field Programmable Gate Arrays  \nFPR False Positive Rate  \nGNN Graph Neural Networks  \nGPU Graphics processing unit HDL Hardware Description Language HEP High Energy Physics  \nHLS High-Level Synthesis  \nHLS4ML High-Level Synthesis for Machine Learning HL-LHC High-Luminosity upgrade  \nIN  \nL1T  \nLHC  \nLSTM  \nML  \nMLP  \nNLP  \nNN  \nPT  \nROC  \nRTL  \nRNN  \nTPR  \n5  \nInterac","cbCaidCjUi4HwdwP","https://ap.wps.com/l/cbCaidCjUi4HwdwP","pdf",1518099,1,60,"English","en",105,"# 1 INTRODUCTION\n## 1.1 Background\n## 1.2 Objectives and delimitations\n## 1.3 Structure of the thesis\n# 2 PARTICLE PHYSICS RESEARCH WITH MACHINE LEARNING\n## 2.1 Boosted Decision Trees\n## 2.2 Deep Neural Networks\n## 2.3 Graph Neural Networks\n## 2.4 Transformer Neural Networks\n## 2.5 Machine Learned Likelihoods\n## 2.6 Real-time Tracking and Triggering in High-Energy Physics\n## 2.7 Summary\n# 3 PARTICLE JET CLASSIFICATION BY MACHINE LEARNING ON FIELD-PROGRAMMABLE GATE ARRAY\n## 3.1 Proposed Methodology\n## 3.2 Model Training and Deployment\n## 3.3 Deep Sets Model on Field-Programmable Gate Array\n# 4 EXPERIMENTS\n## 4.1 Data\n## 4.2 Evaluation criteria\n## 4.3 Experimental Results\n# 5 DISCUSSION\n## 5.1 Current study\n## 5.2 Future work\n# 6 CONCLUSION\n# REFERENCES","[{\"question\":\"What is the thesis focus for particle jet classification?\",\"answer\":\"The thesis focuses on machine learning models that classify particle jets using Large Hadron Collider (LHC) data, targeting feasibility under HL-LHC computational demands.\"},{\"question\":\"Why is a Deep Sets-based model used in the approach?\",\"answer\":\"Deep Sets is used because jet constituents are unordered and their order can vary in the CMS experiment, which matches the Deep Sets design for permutation-invariant inputs.\"},{\"question\":\"How are FPGA deployment constraints addressed during training?\",\"answer\":\"The model is quantized and pruned during training so it fits FPGA-based hardware constraints, enabling an efficient path toward deployment.\"}]","Particle Jet Classification Using Edge Machine Learning - Master’s Thesis 2024 | PDF",1785722928,151,{"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},"particle-jet-classification-using-edge-machine-learning-masters-thesis-2024","",{"@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/particle-jet-classification-using-edge-machine-learning-masters-thesis-2024/119175/",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 is the thesis focus for particle jet classification?","Question",{"text":75,"@type":76},"The thesis focuses on machine learning models that classify particle jets using Large Hadron Collider (LHC) data, targeting feasibility under HL-LHC computational demands.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why is a Deep Sets-based model used in the approach?",{"text":80,"@type":76},"Deep Sets is used because jet constituents are unordered and their order can vary in the CMS experiment, which matches the Deep Sets design for permutation-invariant inputs.",{"name":82,"@type":73,"acceptedAnswer":83},"How are FPGA deployment constraints addressed during training?",{"text":84,"@type":76},"The model is quantized and pruned during training so it fits FPGA-based hardware constraints, enabling an efficient path toward deployment.","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},[93,97,101,105,109,114,119,122,127,130,134],{"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":21,"slug":108},5,"Comic","comic",{"id":110,"doc_module":4,"doc_module_name":46,"category_name":111,"show_sort_weight":112,"slug":113},6,"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"]