Abstract

Convergence Time Minimization of Federated Learning over Wireless Networks, a paper written by Mingzhe Chen, a postdoctoral researcher jointly trained by Prof. Shuguang Cui's team at CUHK-Shenzhen, and Princeton University, received the Best Paper Award at the 2020 IEEE International Conference on Communications (ICC).

IEEE ICC 

The IEEE International Conference on Communications (ICC) is one of the IEEE Communications Society’s two flagship conferences dedicated to driving innovation in nearly every aspect of communications. Each year, around 3,000 researchers submit proposals for paper presentations and program sessions to be held at the annual conference. After extensive peer review, the best of the proposals is selected for the conference’s Best Paper Awards.

Awarded Paper

In this paper, the convergence time of federated learning (FL), when deployed over a realistic wireless network, is studied. Due to the limited number of resource blocks (RBs) in a wireless network, only a subset of users can be selected to transmit their local FL model parameters to the BS at each learning step. To solve this problem, a probabilistic user selection scheme is proposed. Artificial neural networks (ANNs) are employed, for the first time, to estimate the local FL models of the users, thereby increasing the number of FL users and further reducing the FL convergence time.

In writing the paper, Dr. Mingzhe Chen, among other authors, encountered two main challenges. The first was how to train the neural network and estimate the local FL model parameters of the users without adding any user overhead. It was found that the probabilistic user selection scheme not only optimizes the FL convergence performance, but also provides training samples for the neural network, which were previously unavailable. The second conundrum was how to define the output of the neural network. Initially, the team took the FL model parameters of the users for future moments directly as the output, but the results were not satisfactory. Subsequently, the team identified the difference between two users' models as a better output alternative. Eventually, the paper stood out and won the Best Paper Award at IEEE ICC 2020.

About the Author

Dr. Mingzhe Chen

Postdoctoral fellow with Future Network of Intelligence Institute (CUHK-Shenzhen) and Princeton University

Mingzhe Chen received his Ph.D. degree from the Beijing University of Posts and Telecommunications, Beijing, China, in 2019. He is currently a postdoctoral researcher with the team of Professor Shuguang Cui (CUHK-Shenzhen), and the Department of Electrical Engineering, Princeton University. He was an Exemplary Reviewer of the IEEE Transactions on Wireless Communications and the IEEE Transactions on Communications in 2018 and 2019. He has served as the Co-Chair for the IEEE International Conference on Communications (ICC) Workshop on Edge Machine Learning for 5G Mobile Networks and Beyond (2019). He will soon serve as the Co-Chair for the 2020 GLOBECOM Edge Learning over 5G Networks and Beyond, and as a Guest Editor of the IEEE Journal on Selected Areas in Communications (JSAC). Doctor Chen’s research interests include machine learning, virtual reality, drones, game theory, wireless networks, and cache storage.

About the Professor

Prof. Shuguang CUI

Acting Dean of SSE、Dean of FNii

Shuguang Cui received his Ph.D in Electrical Engineering from Stanford University, California, USA, in 2005. Afterwards, he has been working as assistant, associate, full, Chair Professors in Electrical and Computer Engineering at the Univ. of Arizona, Texas A&M University, and UC Davis, respectively. He is currently the Chair Professor and Acting Dean, the School of Science and Engineering at the Chinese University of Hong Kong (Shenzhen), and the Executive Vice Director of Shenzhen Research Institute of Big Data.

His current research interests focus on data driven large-scale system control and resource management, large data set analysis, IoT system design, energy harvesting based communication system design, and cognitive network optimization. He was selected as the Thomson Reuters Highly Cited Researcher and listed in the Worlds’ Most Influential Scientific Minds by ScienceWatch in 2014. He was the recipient of the IEEE Signal Processing Society 2012 Best Paper Award. He has served as the general co-chair and TPC co-chairs for many IEEE conferences. He has also been serving as the area editor for IEEE Signal Processing Magazine, and associate editors for IEEE Transactions on Big Data, IEEE Transactions on Signal Processing, IEEE JSAC Series on Green Communications and Networking, and IEEE Transactions on Wireless Communications. He has been the elected member for IEEE Signal Processing Society SPCOM Technical Committee (2009~2014) and the elected Chair for IEEE ComSoc Wireless Technical Committee (2017~2018). He is a member of the Steering Committee for both IEEE Transactions on Big Data and IEEE Transactions on Cognitive Communications and Networking. He is also a member of the IEEE ComSoc Emerging Technology Committee. He was elected as an IEEE Fellow in 2013 and an IEEE ComSoc Distinguished Lecturer in 2014.