Over-the-Air Federated Learning via Second-Order Optimization

被引:23
作者
Yang, Peng [1 ,2 ]
Jiang, Yuning [3 ]
Wang, Ting [1 ,2 ]
Zhou, Yong [4 ]
Shi, Yuanming [4 ]
Jones, Colin N. [3 ]
机构
[1] East China Normal Univ, Shanghai Key Lab Trustworthy Comp, Shanghai 200062, Peoples R China
[2] East China Normal Univ, Engn Res Ctr Software Hardware Codesign Technol &, Minist Educ, Shanghai 200062, Peoples R China
[3] Ecole Polytech Fed Lausanne EPFL, Automatic Control Lab, CH-1015 Lausanne, Switzerland
[4] ShanghaiTech Univ, Sch Informat Sci & Technol, Shanghai 201210, Peoples R China
基金
上海市自然科学基金; 瑞士国家科学基金会;
关键词
Federated learning; over-the-air computation; second-order optimization method; COMPUTATION; PRIVACY;
D O I
10.1109/TWC.2022.3185156
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Federated learning (FL) is a promising learning paradigm that can tackle the increasingly prominent isolated data islands problem while keeping users' data locally with privacy and security guarantees. However, FL could result in task-oriented data traffic flows over wireless networks with limited radio resources. To design communication-efficient FL, most of the existing studies employ the first-order federated optimization approach that has a slow convergence rate. This however results in excessive communication rounds for local model updates between the edge devices and edge server. To address this issue, in this paper, we instead propose a novel over-the-air second-order federated optimization algorithm to simultaneously reduce the communication rounds and enable low-latency global model aggregation. This is achieved by exploiting the waveform superposition property of a multi-access channel to implement the distributed second-order optimization algorithm over wireless networks. The convergence behavior of the proposed algorithm is further characterized, which reveals a linear-quadratic convergence rate with an accumulative error term in each iteration. We thus propose a system optimization approach to minimize the accumulated error gap by joint device selection and beamforming design. Numerical results demonstrate the system and communication efficiency compared with the state-of-the-art approaches.
引用
收藏
页码:10560 / 10575
页数:16
相关论文
共 63 条
[1]   Federated Learning Over Wireless Fading Channels [J].
Amiri, Mohammad Mohammadi ;
Gunduz, Deniz .
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, 2020, 19 (05) :3546-3557
[2]   Machine Learning at the Wireless Edge: Distributed Stochastic Gradient Descent Over-the-Air [J].
Amiri, Mohammad Mohammadi ;
Gunduz, Deniz .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2020, 68 (68) :2155-2169
[3]  
[Anonymous], 2013, Markov Chains: Gibbs Fields, Monte Carlo Simulation, and Queues
[4]  
Bernstein J, 2018, PR MACH LEARN RES, V80
[5]  
Bischoff S, 2021, Arxiv, DOI arXiv:2109.02388
[6]  
Bonawitz K, 2019, Arxiv, DOI arXiv:1902.01046
[7]  
Bottou Leon, 2012, Neural Networks: Tricks of the Trade. Second Edition: LNCS 7700, P421, DOI 10.1007/978-3-642-35289-8_25
[8]  
Cao Xiaowen, 2021, arXiv
[9]  
Chang WT, 2020, Arxiv, DOI arXiv:2001.08737
[10]   A Uniform-Forcing Transceiver Design for Over-the-Air Function Computation [J].
Chen, Li ;
Qin, Xiaowei ;
Wei, Guo .
IEEE WIRELESS COMMUNICATIONS LETTERS, 2018, 7 (06) :942-945