Communication Dans Un Congrès Année : 2022

Asynchronous Decentralized Learning over Unreliable Wireless Networks

Eunjeong Jeong
Matteo Zecchin
Marios Kountouris

Résumé

Decentralized learning enables edge users to collaboratively train models by exchanging information via deviceto-device communication, yet prior works have been limited to wireless networks with fixed topologies and reliable workers. In this work, we propose an asynchronous decentralized stochastic gradient descent (DSGD) algorithm, which is robust to the inherent computation and communication failures occurring at the wireless network edge. We theoretically analyze its performance and establish a non-asymptotic convergence guarantee. Experimental results corroborate our analysis, demonstrating the benefits of asynchronicity and outdated gradient information reuse in decentralized learning over unreliable wireless networks.

Fichier principal
Vignette du fichier
publi-6800.pdf (646.18 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04143164 , version 1 (27-06-2023)

Licence

Identifiants

Citer

Eunjeong Jeong, Matteo Zecchin, Marios Kountouris. Asynchronous Decentralized Learning over Unreliable Wireless Networks. ICC 2022, IEEE International Conference on Communications, IEEE, May 2022, Seoul, South Korea. pp.607-612, ⟨10.1109/ICC45855.2022.9838891⟩. ⟨hal-04143164⟩

Collections

106 Consultations
110 Téléchargements

Altmetric

Partager

  • More