An Accelerated Method For Decentralized Distributed Stochastic Optimization Over Time-Varying Graphs

被引:5
|
作者
Rogozin, Alexander [1 ,2 ]
Bochko, Mikhail [1 ,2 ]
Dvurechensky, Pavel [3 ,4 ]
Gasnikov, Alexander [5 ]
Lukoshkin, Vladislav [6 ]
机构
[1] Moscow Inst Phys & Technol, Moscow, Russia
[2] HSE Univ, Moscow, Russia
[3] Weierstrass Inst Appl Anal & Stochast, Berlin, Germany
[4] HSE Univ, Inst Informat Transmiss Problems RAS, Moscow, Russia
[5] HSE Univ, Inst Informat Transmiss Problems RAS, Moscow Inst Phys & Technol, Moscow, Russia
[6] Skolkovo Inst Sci & Technol, Moscow, Russia
关键词
stochastic optimization; decentralized distributed optimization; time-varying network; INTERMEDIATE GRADIENT-METHOD; ASYMPTOTIC AGREEMENT; CONVEX; ALGORITHMS; CONVERGENCE;
D O I
10.1109/CDC45484.2021.9683110
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We consider a distributed stochastic optimization problem that is solved by a decentralized network of agents with only local communication between neighboring agents. The goal of the whole system is to minimize a global objective function given as a sum of local objectives held by each agent. Each local objective is defined as an expectation of a convex smooth random function and the agent is allowed to sample stochastic gradients for this function. For this setting we propose the first accelerated (in the sense of Nesterov's acceleration) method that simultaneously attains optimal up to a logarithmic factor communication and oracle complexity bounds for smooth strongly convex distributed stochastic optimization. We also consider the case when the communication graph is allowed to vary with time and obtain complexity bounds for our algorithm, which are the first upper complexity bounds for this setting in the literature.
引用
收藏
页码:3367 / 3373
页数:7
相关论文
共 50 条
  • [1] Accelerated Gradient Tracking over Time-varying Graphs for Decentralized Optimization
    Li, Huan
    Lin, Zhouchen
    JOURNAL OF MACHINE LEARNING RESEARCH, 2024, 25
  • [2] Accelerated Distributed Stochastic Nonconvex Optimization Over Time-Varying Directed Networks
    Chen, Yiyue
    Hashemi, Abolfazl
    Vikalo, Haris
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2025, 70 (04) : 2196 - 2211
  • [3] A Geometrically Convergent Method for Distributed Optimization over Time-Varying Graphs
    Nedich, Angelia
    Olshevsky, Alex
    Shi, Wei
    2016 IEEE 55TH CONFERENCE ON DECISION AND CONTROL (CDC), 2016, : 1023 - 1029
  • [4] Distributed Optimization Over Time-Varying Directed Graphs
    Nedic, Angelia
    Olshevsky, Alex
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2015, 60 (03) : 601 - 615
  • [5] Distributed optimization over time-varying directed graphs
    Nedic, Angelia
    Olshevsky, Alex
    2013 IEEE 52ND ANNUAL CONFERENCE ON DECISION AND CONTROL (CDC), 2013, : 6855 - 6860
  • [6] ADOM: Accelerated Decentralized Optimization Method for Time-Varying Networks
    Kovalev, Dmitry
    Shulgin, Egor
    Richtarik, Peter
    Rogozin, Alexander
    Gasnikov, Alexander
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 139, 2021, 139
  • [7] Decentralized Optimization Over Time-Varying Directed Graphs With Row and Column-Stochastic Matrices
    Saadatniaki, Fakhteh
    Xin, Ran
    Khan, Usman A.
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2020, 65 (11) : 4769 - 4780
  • [8] Distributed Stochastic Algorithm for Convex Optimization Problem Over Time-varying Directed Graphs
    Cheng, Songsong
    Wang, Yinghui
    Hong, Yiguang
    2019 IEEE 15TH INTERNATIONAL CONFERENCE ON CONTROL AND AUTOMATION (ICCA), 2019, : 476 - 481
  • [9] A Geometrically Converging Dual Method for Distributed Optimization Over Time-Varying Graphs
    Maros, Marie
    Jalden, Joakim
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2021, 66 (06) : 2465 - 2479
  • [10] An accelerated distributed optimization algorithm over time-varying digraphs with column-stochastic matrices
    Shi, Xiasheng
    Liu, Hanlin
    Chen, Jiahao
    Wang, Xuesong
    PROCEEDINGS OF THE 33RD CHINESE CONTROL AND DECISION CONFERENCE (CCDC 2021), 2021, : 2124 - 2129