ZTE

ZTE Communications ›› 2020, Vol. 18 ›› Issue (2): 20-30.DOI: 10.12142/ZTECOM.202002004

• Special Topic • Previous Articles     Next Articles

Joint User Selection and Resource Allocation for Fast Federated Edge Learning

JIANG Zhihui(), HE Yinghui, YU Guanding   

  1. College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, Zhejiang 310027, China
  • Received:2020-01-31 Online:2020-06-25 Published:2020-08-07
  • About author:JIANG Zhihui （zhihui.jiang@zju.edu.cn） received the B.E. degree in information engineering from Zhejiang University, China in 2020, where she is currently pursuing the master’s degree with the College of Information Science and Electronic Engineering. Her research interests mainly include federated learning and edge learning.|HE Yinghui received the B.E. degree in information engineering from Zhejiang University, China in 2018, where he is currently pursuing the master’s degree with the College of Information Science and Electronic Engineering. His research interests mainly include mobile edge computing and device-to-device communications.|y. He joined Zhejiang University, in 2006, where he is currently a full professor with the College of Information and Electronic Engineering. From 2013 to 2015, he was a visiting professor with the School of Electrical and Computer Engineering, Georgia Institute of Technology, USA. His research interests include 5G communications and networks, mobile edge computing, and machine learning for wireless networks. Dr. YU received the 2016 IEEE ComSoc Asia-Pacific Outstanding Young Researcher Award. He regularly chairs the technical program committee boards of prominent IEEE conferences, such as ICC, GLOBECOM, and VTC. He also serves as a symposium co-chair for IEEE Globecom 2019 and the track chair for IEEE VTC 2019’ Fall. He has served as a guest editor for the IEEE Communications Magazine special issue on full-duplex communications, an editor for the IEEE Journal on Selected Areas in Communications Series on green communications and networking, a leading guest editor for the IEEE Wireless Communications Magazine special issue on LTE in unlicensed spectrum, and an editor for the IEEE Access. He serves as an editor for the IEEE Transactions on Green Communications and Networking and the IEEE Wireless Communications Letters.|YU Guanding received the B.E. and Ph.D. degrees in communication engineering from Zhejiang University, China in 2001 and 2006, respectivel
  • Supported by:
    the National Natural Science Foundation of China(61671407)

RichHTML

PDF

Abstract:

By periodically aggregating local learning updates from edge users, federated edge learning (FEEL) is envisioned as a promising means to reap the benefit of local rich data and protect users’ privacy. However, the scarce wireless communication resource greatly limits the number of participated users and is regarded as the main bottleneck which hinders the development of FEEL. To tackle this issue, we propose a user selection policy based on data importance for FEEL system. In order to quantify the data importance of each user, we first analyze the relationship between the loss decay and the squared norm of gradient. Then, we formulate a combinatorial optimization problem to maximize the learning efficiency by jointly considering user selection and communication resource allocation. By problem transformation and relaxation, the optimal user selection policy and resource allocation are derived, and a polynomial-time optimal algorithm is developed. Finally, we deploy two commonly used deep neural network (DNN) models for simulation. The results validate that our proposed algorithm has strong generalization ability and can attain higher learning efficiency compared with other traditional algorithms.

Key words: data importance, federated edge learning, learning accuracy, learning efficiency, resource allocation, user selection