ZTE Communications ›› 2023, Vol. 21 ›› Issue (1): 15-24.DOI: 10.12142/ZTECOM.202301003

• Special Topic • Previous Articles     Next Articles

Reliable and Privacy-Preserving Federated Learning with Anomalous Users

ZHANG Weiting1, LIANG Haotian2, XU Yuhua2, ZHANG Chuan2()   

  1. 1.Beijing Jiaotong University, Beijing 100091, China
    2.Beijing Institute of Technology, Beijing 100081, China
  • Received:2022-11-01 Online:2023-03-25 Published:2023-03-22
  • Contact: ZHANG Chuan
  • About author:ZHANG Weiting received his PhD degree in communication and information systems from Beijing Jiaotong University, China in 2021. From Nov. 2019 to Nov. 2020, he was a visiting PhD student with the BBCR Group, Department of Electrical and Computer Engineering, University of Waterloo, Canada. He is currently an associate professor with the School of Electronic and Information Engineering, Beijing Jiaotong University. His research interests include industrial Internet of Things, edge intelligence, and machine learning for wireless networks.|LIANG Haotian received his BS degree from Lanzhou University, China in 2022. He is currently working towards his master’s degree in the School of Cyberspace Science and Technology, Beijing Institute of Technology, China. His research interests include machine learning security, Internet of Things security, and cloud security.|XU Yuhua is currently an undergraduate student in School of Computer Science and Technology, Beijing Institute of Technology, China. She is currently working at the research laboratory of advanced network and data security at the School of Cyberspace Science and Technology, Beijing Institute of Technology. Her research interests include applied cryptography and blockchain.|ZHANG Chuan(chuanz@bit.edu.cn) received his PhD degree in computer science from Beijing Institute of Technology, China in 2021. From Sept. 2019 to Sept. 2020, he worked as a visiting PhD student with the BBCR Group, Department of Electrical and Computer Engineering, University of Waterloo, Canada. He is currently an assistant professor at the School of Cyberspace Science and Technology, Beijing Institute of Technology, China. His research interests include secure data services in cloud computing, applied cryptography, machine learning, and blockchain.
  • Supported by:
    the Fundamental Research Funds for Central Universities(2022RC006);the National Natural Science Foundation of China(62201029);the BIT Research and Innovation Promoting Project(2022YCXZ031);the Shandong Provincial Key Research and Development Program(2021CXGC010106);the China Postdoctoral Science Foundation(2021M700435)

Abstract:

Recently, various privacy-preserving schemes have been proposed to resolve privacy issues in federated learning (FL). However, most of them ignore the fact that anomalous users holding low-quality data may reduce the accuracy of trained models. Although some existing works manage to solve this problem, they either lack privacy protection for users’ sensitive information or introduce a two-cloud model that is difficult to find in reality. A reliable and privacy-preserving FL scheme named reliable and privacy-preserving federated learning (RPPFL) based on a single-cloud model is proposed. Specifically, inspired by the truth discovery technique, we design an approach to identify the user’s reliability and thereby decrease the impact of anomalous users. In addition, an additively homomorphic cryptosystem is utilized to provide comprehensive privacy preservation (user’s local gradient privacy and reliability privacy). We give rigorous theoretical analysis to show the security of RPPFL. Based on open datasets, we conduct extensive experiments to demonstrate that RPPEL compares favorably with existing works in terms of efficiency and accuracy.

Key words: federated learning, anomalous user, privacy preservation, reliability, homomorphic cryptosystem