Faster Convergence With Less Communication: Broadcast-Based Subgraph Sampling for Decentralized Learning Over Wireless Networks

Decentralized stochastic gradient descent (D-SGD) is a widely adopted optimization algorithm for decentralized training of machine learning models across networked agents. A crucial part of D-SGD is the consensus-based model averaging, which heavily relies on information exchange and fusion among th...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE open journal of the Communications Society Vol. 6; pp. 1497 - 1511
Main Authors:	Perez Herrera, Daniel, Chen, Zheng, Larsson, Erik G.
Format:	Journal Article
Language:	English
Published:	New York IEEE 2025 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:	Algorithms broadcast Broadcasting Channels Collision avoidance Communication Convergence Costs D-SGD Decentralized machine learning Graph theory Laplace equations Machine learning Network topology node scheduling Nodes Optimization Sampling methods Scheduling Stochastic processes Symmetric matrices Topology Training Vectors Wireless networks broadcast Costs Symmetric matrices Laplace equations Stochastic processes wireless networks Vectors Topology node scheduling Optimization Convergence Training Network topology D-SGD Decentralized machine learning
ISSN:	2644-125X, 2644-125X
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Be the first to leave a comment!