[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118216-en":3,"doc-seo-118216-105":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},118216,8796095461610,"Oliver","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Towards Plausible Collaborative Machine Learning - Privacy, Efficiency and Fairness","This dissertation tackles privacy, communication efficiency, and fairness issues arising in collaborative distributed machine learning. As decentralized edge learning relies on local private samples and frequent parameter exchanges, it creates strong privacy exposure and communication bottlenecks. Model outputs can also introduce discriminatory outcomes when sensitive attributes shape population groups. The work proposes differentially private ADMM methods with Gaussian noise using decaying variance and noisy approximate solutions, and a decentralized gradient descent scheme using DP noise plus random quantization to improve communication efficiency. It further develops functional mechanisms and decision boundary covariance to quantify decision boundary fairness.","Towards Plausible Collaborative Machine Learning: Privacy,  \nEfficiency and Fairness  \nby  \nJiahao Ding  \nA dissertation submitted to the Department of Electrical and Computer Engineering,  \nCullen College of Engineering  \nin partial fulfillment of the requirements for the degree of  \nDoctor of Philosophy  \nin Electrical Engineering  \nChair of Committee: Miao Pan  \nCommittee Member: Zhu Han  \nCommittee Member: Hien Van Nguyen  \nCommittee Member: Xin Fu  \nCommittee Member: Yanmin Gong  \nUniversity of Houston  \nMay 2022  \nCopyright 2022, Jiahao Ding  \nACKNOWLEDGMENTS  \nI have been fortunate to complete this dissertation with plenty of help and support from amazing mentors, colleagues, collaborators, and friends. I hereby express my sincere gratitude to all of them. First and foremost, I would like to thank my advisor, Dr. Miao Pan, for his keen insights and plentiful encouragements, for giving me tremendous support, and for giving me a chance as a beginner in the field and nurtured me to become a successful researcher. It has been a great privilege and honor to work and study under his guidance.  \nI would like to express my deepest appreciation to my dissertation committee members Dr. Zhu Han, Dr. Hien Van Nguyen, Dr. Xin Fu, and Dr. Yanmin Gong, for helpful discussions, the ingenious suggestions and insightful feedback. Their constructive comments significantly improve the quality of this dissertation. I am also grateful to all other researchers I have collaborated with: Dr. Guannan Liang, Dr. Jinbo Bi, Dr. Di Wang, Dr. Xiaohuan Li, Dr. Junyi Wang, Dr. Maoqiang Wu, Dr. Rong Yu, Dr. Mingsong Chen, Dr. Kaiping Xue, Dr. Chi Zhang, Dr. Haijun Zhang, Dr. Yuanxiong Guo, Dr. Haixia Zhang, Dr. Dongfeng Yuan, and Tian Liu. I have been truly honored to work with these excellent researchers.  \nMy gratitude also goes to all of my friends and colleagues in the AI, Networking Technologies and Security Laboratory (ANTS Lab) at UH ECE department, Dr. Jingyi Wang, Dr. Sai Mounika Errapotu, Dr. Debing Wei, Dr. Xinyue Zhang, Dian Shi, Pavana Prakash, Rui Chen, Chenpei Huang, and many others. It has been wonderful to meet and work with you in Houston.  \nFinally, I would like to thank my parents for their unconditional love and support. Thank you for having my back and always believing in me. None of my achievements would be possible without you. This dissertation is dedicated to them.  \nABSTRACT  \nNowadays, the development of machine learning shows great potential in a variety of fields, such as retail, healthcare, and insurance. Effective machine learning models can automatically learn useful information from a large amount of data and provide decisions with high average accuracy. Although machine learning has infiltrated into many areas due to its advantages, a vast amount of data has been generated at an ever-increasing rate, which leads to significant computational complexity for data collection and processing via a centralized machine learning approach. Distributed machine learning thus has received huge interest due to its capability of exploiting the collective computing power of edge devices. However, during the learning process, model updates using local private samples and large-scale parameter exchanges among agents impose severe privacy concerns and communication bottlenecks. Moreover, the decisions and predictions offered by the learning models may cause certain fairness concerns among population groups of interest, when the grouping is based on such sensitive attributes as race and gender.  \nTo address those challenges, in this dissertation, we first propose a number of differentially private Alternating Direction Method of Multipliers (ADMM) algorithms that leverage two key ideas to balance the privacy-accuracy tradeoff: (1) adding Gaussian noise with decaying variance to reduce the negative effects of noise addition and maintain the convergence behaviors; and (2) outputting a noisy approximate solution for the perturbed objective t","cbCaiomYux7z9rtO","https://ap.wps.com/l/cbCaiomYux7z9rtO","pdf",1632116,1,158,"English","en",105,"# Introduction\n## Overview of Dissertation Contributions and Structure\n# Preliminaries\n# Plausible Differently Private ADMM Based Distributed Machine Learning\n## Differentially Private Robust ADMM\n## Plausible Private ADMM\n## Improved Plausible Private ADMM\n## Omitted Proofs\n# Differentially Private and Communication Efficient Decentralized Gradient Descent\n## Related Work\n## Problem Setting and Preliminaries\n## Main Methods","[{\"question\":\"What core problems does the dissertation address in collaborative distributed machine learning?\",\"answer\":\"It addresses privacy risks from local private samples and parameter exchanges, communication bottlenecks in distributed learning, and fairness concerns where predictions may discriminate across groups defined by sensitive attributes.\"},{\"question\":\"How do the proposed differentially private ADMM algorithms improve the privacy-accuracy tradeoff?\",\"answer\":\"They balance privacy and accuracy by adding Gaussian noise with decaying variance to preserve convergence behavior and by releasing a noisy approximate solution to avoid being constrained by exact optimal solutions at each ADMM iteration.\"},{\"question\":\"How is communication efficiency enforced alongside differential privacy in decentralized gradient descent?\",\"answer\":\"Local model updates integrate DP noise together with a random quantization operator, so differential privacy and communication efficiency are enforced simultaneously.\"}]","Towards Plausible Collaborative Machine Learning - Privacy, Efficiency and Fairness | PDF",1785682314,398,{"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},"towards-plausible-collaborative-machine-learning-privacy-efficiency-and-fairness","",{"@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/towards-plausible-collaborative-machine-learning-privacy-efficiency-and-fairness/118216/",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","2026-08-02",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 core problems does the dissertation address in collaborative distributed machine learning?","Question",{"text":75,"@type":76},"It addresses privacy risks from local private samples and parameter exchanges, communication bottlenecks in distributed learning, and fairness concerns where predictions may discriminate across groups defined by sensitive attributes.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How do the proposed differentially private ADMM algorithms improve the privacy-accuracy tradeoff?",{"text":80,"@type":76},"They balance privacy and accuracy by adding Gaussian noise with decaying variance to preserve convergence behavior and by releasing a noisy approximate solution to avoid being constrained by exact optimal solutions at each ADMM iteration.",{"name":82,"@type":73,"acceptedAnswer":83},"How is communication efficiency enforced alongside differential privacy in decentralized gradient descent?",{"text":84,"@type":76},"Local model updates integrate DP noise together with a random quantization operator, so differential privacy and communication efficiency are enforced simultaneously.","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"]