[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120112-en":3,"doc-seo-120112-105":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":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},120112,8796095462418,"Noah","https://ap-avatar.wpscdn.com/avatar/80000253c1241d02b47?x-image-process=image/resize,m_fixed,w_180,h_180&k=1778826106357471780",8,"Research & Report","Multi-objective hyperparameter optimisation for edge machine learning - Master’s thesis 2024","The high-luminosity upgrade of the Large Hadron Collider increases the need for fast, correct trigger decisions for classifying particle jets, motivating advanced edge computing. Field-Programmable Gate Arrays can meet strict latency constraints but offer limited resources, while typical machine learning models require high memory and computation that are difficult to deploy directly. This thesis compares popular multi-objective hyperparameter optimisation methods and libraries to select an approach for edge deployment under hardware limits, and designs two task-specific loss functions to achieve strong model performance.","MULTI-OBJECTIVE HYPERPARAMETER OPTIMISATION FOR EDGE MACHINE LEARNING  \nLappeenranta-Lahti University of Technology LUT  \nMaster’s Program in Computational Engineering, Master’s Thesis 2024  \nTing Wang  \nExaminers: Professor Lasse Lensu  \nD.Sc. (Tech.) Henri Petrow  \nABSTRACT  \nLappeenranta-Lahti University of Technology LUT School of Engineering Science  \nComputational Engineering  \nTing Wang  \nMulti-objective hyperparameter optimisation for edge machine learning  \nMaster’s thesis  \n2024  \n56 pages, 15 figures, 13 tables  \nExaminers: Professor Lasse Lensu and D.Sc. (Tech.) Henri Petrow  \nKeywords: multi-objective hyperparameter optimisation, field-programmable gate array, edge computing, large hadron collider, compact muon solenoid trigger, machine learning, loss function  \nThe high-luminosity upgrade of the Large Hadron Collider at Cern intensifies the demand for advanced trigger algorithms that provide rapid and correct decision-making for classifying particle jets. This challenge can be solved by applying high-performance edge computing. Field-Programmable Gate Array as a form of edge computing hardware can meet strict latency requirements, but they have limited resources for the implementations. Machine Learning models are typically large in size and suffer from high memory and computational requirements, making them difficult to exploit directly on Field-Programmable Gate Arrays. To obtain a high-performance model that meets the hardware-related restrictions, hyperparameter optimisation has an important role. To conduct multi-objective hyperparameter optimisation for edge machine learning based on the two aforementioned issues, this thesis compares the current popular hyperparameter optimisation methods and libraries to appropriately select one for multi-objective tasks. By conducting multiple hyperparameter optimisation experiments, two new loss functions were designed to satisfy different multi-objective tasks for achieving best model performance. The experiments with the Tree-structured Parzen Estimator and the Optuna and hls4ml libraries demonstrate that by defining the loss function appropriately, it is feasible to implement multi-objective hyperparameter optimisation.  \nACKNOWLEDGEMENTS  \nThe completion of this thesis also means the end of my master’s studies.  \nI would like to express my appreciation to Henri Petrow, my first supervisor and second examiner, for his concern and constant help and patience throughout my experiment.  \nI would like to express my sincere gratitude to my first examiner and second supervisor, Lasse Lensu, for his positive encouragement, acknowledgement and careful guidance from the beginning to the end of the thesis writing process.  \nA much more special thanks to my boyfriend, Sun, for all his unrelenting support and for encouraging me to come to Finland and start a new chapter in my life.  \nI have to say thank you to my dear wonderful family, especially my grandmother and older sister, Ming, and my younger sister, Xue, for their companionship and motivation. Even though we are thousands of miles apart, Ming and Xue’s love and companionship always made me feel warm and supported me through the difficult hours of the long polar night.  \nThanks to my friend Yue, for being the first to offer me traditional Chinese medicine treatment options and specially tailored remedies every time I felt bad.  \nThanks to my friend Passacholamas Sukolratchai, all the time we spent learning and having fun together during these two years will be a joyful imprint of my life, shining like golden raindrops. I will never forget it. Thank you Jose Mugumya for encouraging meand supporting me when I was confused about my studies.  \nDeclaration of AI use: During the preparation of this master’s thesis, TING WANG, the author of the thesis, used Deepl and Grammarly in order to translate, polish and confirm language logic. After using Deepl and Grammarly, the author reviewed and edited the content and takes full respon","cbCaivauzgwMsBX7","https://ap.wps.com/l/cbCaivauzgwMsBX7","pdf",4007993,1,56,"English","en",105,"# 1 INTRODUCTION\n## 1.1 Background\n## 1.2 Objectives and Delimitation\n## 1.3 Structure of the Thesis\n# 2 Optimisation and Evaluation of Machine Learning Models\n## 2.1 A Brief Introduction to Parameters in Machine Learning Models\n## 2.2 Hyperparameters and their Optimisation\n## 2.3 Hyperparameter Optimisation Methods\n## 2.4 Multi-objective Hyperparameter Optimisation\n## 2.5 Hyperparameter Optimisation Libraries\n## 2.6 Evaluation Metrics\n# 3 Multi-objective Optimisation of Neural Networks for Resource-Restricted Deployment\n## 3.1 Neural Network Architecture\n## 3.2 Convert Neural Network to Field-Programmable Gate Array (FPGA) with hls4ml","[{\"question\":\"What problem does the thesis address for edge machine learning in particle physics triggers?\",\"answer\":\"It targets the need for rapid and accurate trigger decisions for classifying particle jets, where edge computing is required but FPGA resources are limited.\"},{\"question\":\"Why is hyperparameter optimisation important for deploying machine learning on FPGAs?\",\"answer\":\"Machine learning models are often too large and resource-intensive to run directly on FPGAs, so hyperparameter optimisation helps find models that fit hardware restrictions while maintaining performance.\"},{\"question\":\"How does the thesis enable multi-objective hyperparameter optimisation on edge hardware?\",\"answer\":\"It compares hyperparameter optimisation methods and libraries and designs loss functions tailored to different multi-objective tasks, showing feasibility with Tree-structured Parzen Estimator and Optuna/hls4ml workflows.\"}]","Multi-objective hyperparameter optimisation for edge machine learning - Master’s thesis 2024 | PDF",1785728256,141,{"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},"multi-objective-hyperparameter-optimisation-for-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/multi-objective-hyperparameter-optimisation-for-edge-machine-learning-masters-thesis-2024/120112/",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":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the thesis address for edge machine learning in particle physics triggers?","Question",{"text":75,"@type":76},"It targets the need for rapid and accurate trigger decisions for classifying particle jets, where edge computing is required but FPGA resources are limited.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why is hyperparameter optimisation important for deploying machine learning on FPGAs?",{"text":80,"@type":76},"Machine learning models are often too large and resource-intensive to run directly on FPGAs, so hyperparameter optimisation helps find models that fit hardware restrictions while maintaining performance.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the thesis enable multi-objective hyperparameter optimisation on edge hardware?",{"text":84,"@type":76},"It compares hyperparameter optimisation methods and libraries and designs loss functions tailored to different multi-objective tasks, showing feasibility with Tree-structured Parzen Estimator and Optuna/hls4ml workflows.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]