Distributed consensus problem with caching on federated learning framework

Federated learning framework facilitates more applications of deep learning algorithms on the existing network architectures, where the model parameters are aggregated in a centralized manner. However, some of federated learning participants are often inaccessible, such as in a power shortage or dor...

Full description

Saved in:
Bibliographic Details
Main Authors: Xin Yan, Yiming Qin, Xiaodong Hu, Xiaoling Xiao
Format: Article
Language:English
Published: Wiley 2022-04-01
Series:International Journal of Distributed Sensor Networks
Online Access:https://doi.org/10.1177/15501329221092932
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849307788107317248
author Xin Yan
Yiming Qin
Xiaodong Hu
Xiaoling Xiao
author_facet Xin Yan
Yiming Qin
Xiaodong Hu
Xiaoling Xiao
author_sort Xin Yan
collection DOAJ
description Federated learning framework facilitates more applications of deep learning algorithms on the existing network architectures, where the model parameters are aggregated in a centralized manner. However, some of federated learning participants are often inaccessible, such as in a power shortage or dormant state. That will force us to explore the possibility that the parameter aggregation is operated in an ad hoc manner, which is based on consensus computing. On the contrary, since caching mechanism is indispensable to any federated learning mobile node, it is necessary to investigate the connection between it and consensus computing. In this article, we first propose a novel federated learning paradigm, which supports an ad hoc operation mode for federated learning participants. Second, a discrete-time dynamic equation and its control law are formulated to satisfy the demands from federated learning framework, with a quantized caching scheme designed to mask the uncertainties from both asynchronous updates and measurement noises. Then, the consensus conditions and the convergence of the consensus protocol are deduced analytically, and a quantized caching strategy to optimize the convergence speed is provided. Our major contribution is to give the basic theories of distributed consensus problem for federated learning framework, and the theoretical results are validated by numerical simulations.
format Article
id doaj-art-288c50b160bb4d9aa45af80bafb561ad
institution Kabale University
issn 1550-1477
language English
publishDate 2022-04-01
publisher Wiley
record_format Article
series International Journal of Distributed Sensor Networks
spelling doaj-art-288c50b160bb4d9aa45af80bafb561ad2025-08-20T03:54:38ZengWileyInternational Journal of Distributed Sensor Networks1550-14772022-04-011810.1177/15501329221092932Distributed consensus problem with caching on federated learning frameworkXin Yan0Yiming Qin1Xiaodong Hu2Xiaoling Xiao3School of Computer Science and Artificial Intelligence, Wuhan University of Technology, Wuhan, ChinaSchool of Computer Science and Artificial Intelligence, Wuhan University of Technology, Wuhan, ChinaThe Faculty of Education, The University of Hong Kong, Hong Kong, ChinaSchool of Computer Science, Yangtze University, Jingzhou, ChinaFederated learning framework facilitates more applications of deep learning algorithms on the existing network architectures, where the model parameters are aggregated in a centralized manner. However, some of federated learning participants are often inaccessible, such as in a power shortage or dormant state. That will force us to explore the possibility that the parameter aggregation is operated in an ad hoc manner, which is based on consensus computing. On the contrary, since caching mechanism is indispensable to any federated learning mobile node, it is necessary to investigate the connection between it and consensus computing. In this article, we first propose a novel federated learning paradigm, which supports an ad hoc operation mode for federated learning participants. Second, a discrete-time dynamic equation and its control law are formulated to satisfy the demands from federated learning framework, with a quantized caching scheme designed to mask the uncertainties from both asynchronous updates and measurement noises. Then, the consensus conditions and the convergence of the consensus protocol are deduced analytically, and a quantized caching strategy to optimize the convergence speed is provided. Our major contribution is to give the basic theories of distributed consensus problem for federated learning framework, and the theoretical results are validated by numerical simulations.https://doi.org/10.1177/15501329221092932
spellingShingle Xin Yan
Yiming Qin
Xiaodong Hu
Xiaoling Xiao
Distributed consensus problem with caching on federated learning framework
International Journal of Distributed Sensor Networks
title Distributed consensus problem with caching on federated learning framework
title_full Distributed consensus problem with caching on federated learning framework
title_fullStr Distributed consensus problem with caching on federated learning framework
title_full_unstemmed Distributed consensus problem with caching on federated learning framework
title_short Distributed consensus problem with caching on federated learning framework
title_sort distributed consensus problem with caching on federated learning framework
url https://doi.org/10.1177/15501329221092932
work_keys_str_mv AT xinyan distributedconsensusproblemwithcachingonfederatedlearningframework
AT yimingqin distributedconsensusproblemwithcachingonfederatedlearningframework
AT xiaodonghu distributedconsensusproblemwithcachingonfederatedlearningframework
AT xiaolingxiao distributedconsensusproblemwithcachingonfederatedlearningframework