Communication-Efficient Federated Learning for Heterogeneous Edge Devices Based on Adaptive Gradient Quantization

Heting Liu, Fang He, Guohong Cao

Research output: Chapter in Book/Report/Conference proceedingConference contribution

22 Scopus citations

Abstract

Federated learning (FL) enables geographically dispersed edge devices (i.e., clients) to learn a global model without sharing the local datasets, where each client performs gradient descent with its local data and uploads the gradients to a central server to update the global model. However, FL faces massive communication overhead resulted from uploading the gradients in each training round. To address this problem, most existing research compresses the gradients with fixed and unified quantization for all the clients, which neither seeks adaptive quantization due to the varying gradient norms at different rounds, nor exploits the heterogeneity of the clients to accelerate FL. In this paper, we propose a novel adaptive and heterogeneous gradient quantization algorithm (AdaGQ) for FL to minimize the wall-clock training time from two aspects: i) adaptive quantization which exploits the change of gradient norm to adjust the quantization resolution in each training round; and ii) heterogeneous quantization which assigns lower quantization resolution to slow clients to align their training time with other clients to mitigate the communication bottleneck, and higher quantization resolution to fast clients to achieve a better communication efficiency and accuracy tradeoff. Evaluations based on various models and datasets validate the benefits of AdaGQ, reducing the total training time by up to 52.1% compared to baseline algorithms (e.g., FedAvg, QSGD).

Original languageEnglish (US)
Title of host publicationINFOCOM 2023 - IEEE Conference on Computer Communications
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350334142
DOIs
StatePublished - 2023
Event42nd IEEE International Conference on Computer Communications, INFOCOM 2023 - Hybrid, New York City, United States
Duration: May 17 2023May 20 2023

Publication series

NameProceedings - IEEE INFOCOM
Volume2023-May
ISSN (Print)0743-166X

Conference

Conference42nd IEEE International Conference on Computer Communications, INFOCOM 2023
Country/TerritoryUnited States
CityHybrid, New York City
Period5/17/235/20/23

All Science Journal Classification (ASJC) codes

  • General Computer Science
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Communication-Efficient Federated Learning for Heterogeneous Edge Devices Based on Adaptive Gradient Quantization'. Together they form a unique fingerprint.

Cite this