Resource-Constrained Federated Learning: Fundamental Analysis and Optimizations
Date and time: 16 December 2021, 12:00 – 13:00 CET (UTC +1)
Speaker: Chuan Ma, Nanjing University of Science and Technology
Title: Resource-Constrained Federated Learning: Fundamental Analysis and Optimizations
Meeting ID: 695 6088 7455
Watch the recorded presentation:
Abstract: With the rapid development of the Internet-of-Things (IoT), data from intelligent devices is exploding at unprecedented scales. However, conventional centralized ML offers little scalability for efficiently processing such an amount of data. To tackle this challenge, federated learning (FL), which allows decoupling of data provision at clients and aggregating learning models at a centralized server, begins to show its potential advantages and attract increasing attention.
In this talk, we start from the fundamental performance analysis of FL, and conduct a series of studies on resource-constrained settings. Specifically, the whole talk mainly consists of three parts, the differentially private design and unreliable clients of the security issue, the decentralized framework based on blockchain, and the low-latency and resources allocation in a wireless environment. In addition, we also provide some challenges and future directions in FL.
Bio: Dr Chuan Ma received the B.S. degree from the Beijing University of Posts and Telecommunications, Beijing, China, in 2013 and the PhD degree in Telecommunication from the University of Sydney in 2018. He is now working as a lecturer at the School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing, China. He has published more than 20 transaction and conference papers, including a best paper in the WCNC 2018. His research interests include stochastic geometry, device-to-device communication, wireless caching networks and machine learning, and now he is working on big data analysis and privacy preservation in distributed learning.