Distributed Random Reshuffling Over Networks

被引:7
|
作者
Huang, Kun [1 ,2 ]
Li, Xiao [3 ]
Milzarek, Andre [1 ]
Pu, Shi [1 ]
Qiu, Junwen [1 ,2 ]
机构
[1] Chinese Univ Hong Kong, Shenzhen Res Inst Big Data, Sch Data Sci, Shenzhen 518172, Peoples R China
[2] Shenzhen Inst Artificial Intelligence & Robot Soc, Shenzhen 518129, Peoples R China
[3] Chinese Univ Hong Kong, Sch Data Sci, Shenzhen 518172, Peoples R China
关键词
Convergence; Linear programming; Signal processing algorithms; Gradient methods; Distributed databases; Big Data; Machine learning algorithms; Distributed optimization; random reshuffling; stochastic gradient methods; STOCHASTIC OPTIMIZATION; LEARNING-BEHAVIOR; CONVERGENCE; CONSENSUS;
D O I
10.1109/TSP.2023.3262181
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
In this paper, we consider distributed optimization problems where n agents, each possessing a local cost function, collaboratively minimize the average of the local cost functions over a connected network. To solve the problem, we propose a distributed random reshuffling (D-RR) algorithm that invokes the random reshuffling (RR) update in each agent. We show that D-RR inherits favorable characteristics of RR for both smooth strongly convex and smooth nonconvex objective functions. In particular, for smooth strongly convex objective functions, D-RR achieves O(1/T-2) rate of convergence (where T counts the epoch number) in terms of the squared distance between the iterate and the global minimizer. When the objective function is assumed to be smooth nonconvex, we show that D-RR drives the squared norm of the gradient to 0 at a rate of O(1/T-2/3). These convergence results match those of centralized RR (up to constant factors) and outperform the distributed stochastic gradient descent (DSGD) algorithm if we run a relatively large number of epochs. Finally, we conduct a set of numerical experiments to illustrate the efficiency of the proposed D-RR method on both strongly convex and nonconvex distributed optimization problems.
引用
收藏
页码:1143 / 1158
页数:16
相关论文
共 50 条
  • [21] Accelerated AB/Push-Pull Methods for Distributed Optimization Over Time-Varying Directed Networks
    Nguyen, Duong Thuy Anh
    Nguyen, Duong Tung
    Nedic, Angelia
    IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS, 2024, 11 (03): : 1395 - 1407
  • [22] A DISTRIBUTED ALGORITHM FOR DICTIONARY LEARNING OVER NETWORKS
    Zhao, Ming-Min
    Shi, Qingjiang
    Hong, Mingyi
    2016 IEEE GLOBAL CONFERENCE ON SIGNAL AND INFORMATION PROCESSING (GLOBALSIP), 2016, : 505 - 509
  • [23] Distributed Primal-Dual Perturbation Algorithm Over Unbalanced Directed Networks
    Sakuma, Hiroaki
    Hayashi, Naoki
    Takai, Shigemasa
    IEEE ACCESS, 2021, 9 : 75324 - 75335
  • [24] Random Reshuffling: Simple Analysis with Vast Improvements
    Mishchenko, Konstantin
    Khaled, Ahmed
    Richtarik, Peter
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 33, NEURIPS 2020, 2020, 33
  • [25] Sensor networks with random links: Topology design for distributed consensus
    Kar, Soummya
    Moura, Jose M. F.
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2008, 56 (07) : 3315 - 3326
  • [26] Distributed learning for Random Vector Functional-Link networks
    Scardapane, Simone
    Wang, Dianhui
    Panella, Massimo
    Uncini, Aurelio
    INFORMATION SCIENCES, 2015, 301 : 271 - 284
  • [27] Consensus Over Numerosity-Constrained Random Networks
    Abaid, Nicole
    Porfiri, Maurizio
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2011, 56 (03) : 649 - 654
  • [28] Stochastic Proximal Gradient Consensus Over Random Networks
    Hong, Mingyi
    Chang, Tsung-Hui
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2017, 65 (11) : 2933 - 2948
  • [29] Sampled-Data Consensus Over Random Networks
    Wu, Junfeng
    Meng, Ziyang
    Yang, Tao
    Shi, Guodong
    Johansson, Karl Henrik
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2016, 64 (17) : 4479 - 4492
  • [30] Asynchronous Distributed Nonlinear Estimation Over Directed Networks
    Wang, Qianyao
    Yu, Rui
    Meng, Min
    IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING, 2024, 11 (02): : 2062 - 2073