BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Department of Electrical and Computer Engineering (HKUECE) 電機與計算機工程系 - ECPv6.15.20//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Department of Electrical and Computer Engineering (HKUECE) 電機與計算機工程系
X-ORIGINAL-URL:https://ece.hku.hk
X-WR-CALDESC:Events for Department of Electrical and Computer Engineering (HKUECE) 電機與計算機工程系
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:20230101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20240906T153000
DTEND;TZID=Asia/Hong_Kong:20240906T163000
DTSTAMP:20260512T135023
CREATED:20240904T041742Z
LAST-MODIFIED:20250114T042015Z
UID:19089-1725636600-1725640200@ece.hku.hk
SUMMARY:Mechanism Design for Federated Learning with Unstateful Clients
DESCRIPTION:Abstract\nFederated learning (FL) is an attractive distributed machine learning paradigm that enables numerous clients to collaboratively train a model under the coordination of a central server\, while keeping the training data private. However\, without sufficient incentive\, clients may be reluctant to participate in FL due to the associated training costs and variant availability. Moreover\, clients’ availability can be inconsistent\, leading to periodic or random participation—a phenomenon known as unstatefulness. This variability renders existing incentive mechanisms\, designed for full or partial client participation throughout the entire training process\, ineffective. \nIn this talk\, we propose a game-theoretic incentive mechanism for FL with randomized client participation\, where the server adopts a customized pricing strategy to motivate clients to participate at different levels (probabilities). Each client responds to the server’s monetary incentive by choosing its optimal participation level to maximize its profit\, considering both the incurred local cost and its intrinsic value for the model. We show that intrinsic value (internal motivation) introduces the intriguing possibility of bidirectional payments between the server and clients\, leading to a more efficient pricing strategy and enhanced model performance. \nSpeaker\nProf. Bing Luo\nAssistant Professor of Data and Computational Science\,\nDuke Kunshan University (DKU) \nBiography of the Speaker\nProf. Bing Luo is an Assistant Professor of Data and Computational Science at Duke Kunshan University (DKU). He earned his Ph.D. from The University of Melbourne and served as a joint Postdoctoral Researcher at The Chinese University of Hong Kong (Shenzhen) and Yale University. Prior to his Ph.D.\, he worked as a project manager at the China Mobile Corporation Headquarter. His current research interests include the theory and practice of federated and edge learning\, with a focus on optimization and game-theoretical design\, as well as embedded AI for mobile systems. More information can be found in this webpage: https://luobing1008.github.io/ \nOrganiser\nProf. Xianhao Chen\nAssistant Professor\,\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nAll are welcome! We look forward to seeing you!
URL:https://ece.hku.hk/events/20240906-1/
LOCATION:Room CB-603\, 6/F\, Chow Yei Ching Building\, The University of Hong Kong
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/09/1280-2.jpg
END:VEVENT
END:VCALENDAR