Q-GADMM: Quantized group ADMM for communication efficient decentralized machine learning

Elgabli, Anis, Park, Jihong, Bedi, Amrit S., Bennis, Mehdi and Aggarwal, Vaneet 2020, Q-GADMM: Quantized group ADMM for communication efficient decentralized machine learning, in ICASSP 2020 : Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing, Institute of Electrical and Electronics Engineers (IEEE), Piscataway, N.J., pp. 8876-8880, doi: 10.1109/ICASSP40776.2020.9054491.

Attached Files
Name Description MIMEType Size Downloads

Title Q-GADMM: Quantized group ADMM for communication efficient decentralized machine learning
Author(s) Elgabli, Anis
Park, JihongORCID iD for Park, Jihong orcid.org/0000-0001-7623-6552
Bedi, Amrit S.
Bennis, Mehdi
Aggarwal, Vaneet
Conference name ICASSP - IEEE Acoustics, Speech and Signal Processing. International Conference (2020 : Barcelona, Spain)
Conference location Online : Barcelona, Spain
Conference dates 04 - 08 May. 2020
Title of proceedings ICASSP 2020 : Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing
Publication date 2020
Start page 8876
End page 8880
Total pages 5
Publisher Institute of Electrical and Electronics Engineers (IEEE)
Place of publication Piscataway, N.J.
Keyword(s) communication-efficient decentralized machine learning
GADMM
ADMM
quantization
no CORE2020
Summary In this paper, we propose a communication-efficient decen-tralized machine learning (ML) algorithm, coined quantized group ADMM (Q-GADMM). Every worker in Q-GADMM communicates only with two neighbors, and updates its model via the group alternating direct method of multiplier (GADMM), thereby ensuring fast convergence while reducing the number of communication rounds. Furthermore, each worker quantizes its model updates before transmissions, thereby decreasing the communication payload sizes. We prove that Q-GADMM converges to the optimal solution for convex loss functions, and numerically show that Q-GADMM yields 7x less communication cost while achieving almost the same accuracy and convergence speed compared to GADMM without quantization.
ISBN 9781509066315
ISSN 1520-6149
Language eng
DOI 10.1109/ICASSP40776.2020.9054491
Indigenous content off
HERDC Research category E1.1 Full written paper - refereed
Copyright notice ©2020, IEEE
Persistent URL http://hdl.handle.net/10536/DRO/DU:30143315

Connect to link resolver
 
Unless expressly stated otherwise, the copyright for items in DRO is owned by the author, with all rights reserved.

Versions
Version Filter Type
Citation counts: TR Web of Science Citation Count  Cited 0 times in TR Web of Science
Scopus Citation Count Cited 0 times in Scopus
Google Scholar Search Google Scholar
Access Statistics: 9 Abstract Views, 1 File Downloads  -  Detailed Statistics
Created: Mon, 05 Oct 2020, 14:25:06 EST

Every reasonable effort has been made to ensure that permission has been obtained for items included in DRO. If you believe that your rights have been infringed by this repository, please contact drosupport@deakin.edu.au.