[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122808-en":3,"doc-seo-122808-105":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},122808,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Machine Learning Model Splitting on Mobile Edge Networks - Dissertation","Machine learning model splitting on mobile edge networks studies how to partition ML workloads across cellular and edge resources to improve feasibility under bandwidth, latency, and dynamic connectivity constraints. The dissertation develops HiveMind multi-split designs for 5G MEC, including split cost information, multi-objective optimization, and methods for splitting non-linear neural networks. It further proposes an error-tolerant split framework using NeuroMessenger operations to enhance robustness under communication and link-condition changes, and designs omnidirectional millimeter-wave coverage for reliable split execution.","UC San Diego  \nUC San Diego Electronic Theses and Dissertations  \nTitle  \nMachine Learning Model Splitting on Mobile Edge Networks  \nPermalink  \n[https://escholarship.org/uc/item/08q3485r](https://escholarship.org/uc/item/08q3485r)  \nAuthor  \nWang, Song  \nPublication Date  \n2023  \nPeer reviewed|Thesis/dissertation  \n[eScholarship.org](eScholarship.org) Powered by the California Digital Library  \nUniversity of California  \nUNIVERSITY OF CALIFORNIA SAN DIEGO  \nMachine Learning Model Splitting on Mobile Edge Networks  \nA dissertation submitted in partial satisfaction of the requirements for the degree  \nDoctor of Philosophy  \nin  \nElectrical Engineering (Communication Theory and System)  \nby  \nSong Wang  \nCommittee in charge:  \nProfessor Xinyu Zhang, Chair  \nProfessor Sujit Dey  \nProfessor Haojian Jin  \nProfessor Patrick Pannuto  \nProfessor Bhaskar Rao  \nProfessor Ramesh Rao  \n2023  \nCopyright Song Wang, 2023 All rights reserved.  \nThe dissertation of Song Wang is approved, and it is acceptable in quality and form for publication on microfilm and electronically.  \nUniversity of California San Diego  \n2023  \nDEDICATION  \nTo those who silently supported, quietly encouraged, and tirelessly  \nbelieved in this pursuit.  \nEPIGRAPH  \nThe purpose of computing is insight, not numbers.  \n—Richard Hamming  \nTABLE OF CONTENTS  \nDissertation Approval Page ............................... iii  \nDedication ......................................... iv  \nEpigraph ......................................... v  \nTable [of Contents ..................................... vi](of Contents ..................................... vi)  \n[List of Figures ...................................... ix](List of Figures ...................................... ix)  \n[List of Tables ....................................... xii](List of Tables ....................................... xii)  \n[Acknowledgements .................................... xiii](Acknowledgements .................................... xiii)  \n[Vita ............................................ xvi](Vita ............................................ xvi)  \n[Abstract of the Dissertation ............................... xvii](Abstract of the Dissertation ............................... xvii)  \n[Chapter 1 Introduction ................................ 1](Chapter 1 Introduction ................................ 1)  \n1.1 Split ML: Opportunities and Challenges ............. 3  \n1.2 Dissertation Contributions ..................... 6  \nChapter 2 Cellular Native Machine Learning Model Splitting ........... 8  \n2.1 Introduction ............................. 8  \n2.2 Related Work ............................ 13  \n2.2.1 Distributed ML ....................... 13  \n2.2.2 ML model splitting...................... 14  \n2.2.3 Tailoring ML models to edge computing systems..... 15  \n2.3 The Need for Multi-Split in 5G MEC Networks ......... 16  \n2.4 HiveMind Multi-Split Design .................... 19  \n2.4.1 A Primer on 5G MEC for ML ............... 19  \n2.4.2 Problem formulation .................... 21  \n2.4.3 Split Cost Information (SCI) design ............ 24  \n2.4.4 Cost analysis ........................ 31  \n2.4.5 Extension to split DNN training .............. 33  \n2.4.6 Runtime Optimization under Network Dynamics ..... 34  \n2.5 HiveMind Multi-objective Split .................. 37  \n2.6 Splitting Non-Linear Neural Networks ............... 39  \n2.6.1 Split RNN .......................... 40  \n2.6.2 Split Collaborative Learning ................ 41  \n2.7 Evaluation .............................. 44  \n2.7.1 Simulation setup ...................... 44  \n2.7.2 Multi-split performance validation............. 46  \n2.7.3 Performance under network dynamics ........... 52  \n2.7.4 Effectiveness of multi-objective split ............ 54  \n2.7.5 Effectiveness in splitting non-linear ML models ...... 56  \n2.8 Conclusion .............................. 57  \nChapter 3 Error Tolerant ML Model Splitting Over Edge Networks ....... 59  \n3.1 Introduction ","cbCairXg4hRftyy9","https://ap.wps.com/l/cbCairXg4hRftyy9","pdf",9834734,1,181,"English","en",105,"# Chapter 1 Introduction\n## Split ML: Opportunities and Challenges\n## Dissertation Contributions\n# Chapter 2 Cellular Native Machine Learning Model Splitting\n## Introduction\n## Related Work\n## The Need for Multi-Split in 5G MEC Networks\n## HiveMind Multi-Split Design\n## Evaluation\n## Conclusion\n# Chapter 3 Error Tolerant ML Model Splitting Over Edge Networks\n## Introduction\n## Related Work\n## System overview\n## Error-tolerance in Distributed ML\n## NeuroMessenger Operations\n## Evaluation\n## Conclusion\n# Chapter 4 Omnidirectional Millimeter-Wave Coverage for ML Model Splitting\n## Introduction\n## Motivation and Challenges\n## Design\n## Implementation and Experiment setup\n## Evaluation\n## Discussion","[{\"question\":\"What problem does the dissertation address in mobile edge networks?\",\"answer\":\"It addresses how to split machine learning models across edge and cellular resources while coping with opportunities and challenges such as limited communication resources and network dynamics.\"},{\"question\":\"How does HiveMind support multi-split design for 5G MEC?\",\"answer\":\"HiveMind introduces split cost information, cost analysis, runtime optimization under network dynamics, and extensions for split DNN training, together with multi-objective splitting approaches.\"},{\"question\":\"What is the role of error tolerance in distributed split ML?\",\"answer\":\"The dissertation characterizes error tolerance for split ML, compares it with FL settings, and proposes NeuroMessenger operations to enhance robustness against coding and link-condition variations.\"}]","Machine Learning Model Splitting on Mobile Edge Networks - Dissertation | PDF",1785813018,456,{"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},"machine-learning-model-splitting-on-mobile-edge-networks-dissertation","",{"@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/machine-learning-model-splitting-on-mobile-edge-networks-dissertation/122808/",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","2026-08-04",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 problem does the dissertation address in mobile edge networks?","Question",{"text":75,"@type":76},"It addresses how to split machine learning models across edge and cellular resources while coping with opportunities and challenges such as limited communication resources and network dynamics.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does HiveMind support multi-split design for 5G MEC?",{"text":80,"@type":76},"HiveMind introduces split cost information, cost analysis, runtime optimization under network dynamics, and extensions for split DNN training, together with multi-objective splitting approaches.",{"name":82,"@type":73,"acceptedAnswer":83},"What is the role of error tolerance in distributed split ML?",{"text":84,"@type":76},"The dissertation characterizes error tolerance for split ML, compares it with FL settings, and proposes NeuroMessenger operations to enhance robustness against coding and link-condition variations.","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"]