Skip to main navigation Skip to search Skip to main content

Over-the-Air Computation Empowered Federated Learning: A Joint Uplink-Downlink Design

  • Deyou Zhang*
  • , Ming Xiao
  • , Mikael Skoglund
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this paper, we investigate the communication designs of over-the-air computation (AirComp) empowered federated learning (FL) systems considering uplink model aggregation and downlink model dissemination jointly. We first derive an upper bound on the expected difference between the training loss and the optimal loss, which reveals that optimizing the FL performance is equivalent to minimizing the distortion in the received global gradient vector at each edge node. As such, we jointly optimize each edge node transmit and receive equalization coefficients along with the edge server forwarding matrix to minimize the maximum gradient distortion across all edge nodes. We further utilize the MNIST dataset to evaluate the performance of the considered FL system in the context of the handwritten digit recognition task. Experiment results show that deploying multiple antennas at the edge server significantly reduces the distortion in the received global gradient vector, leading to a notable improvement in recognition accuracy compared to the single antenna case.

Original languageEnglish
Title of host publication2023 IEEE 98th Vehicular Technology Conference, VTC 2023-Fall - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350329285
DOIs
StatePublished - 2023
Externally publishedYes
Event98th IEEE Vehicular Technology Conference, VTC 2023-Fall - Hong Kong, China
Duration: 10 Oct 202313 Oct 2023

Publication series

NameIEEE Vehicular Technology Conference
ISSN (Print)1550-2252

Conference

Conference98th IEEE Vehicular Technology Conference, VTC 2023-Fall
Country/TerritoryChina
CityHong Kong
Period10/10/2313/10/23

Keywords

  • Federated learning
  • joint uplink-downlink design
  • over-the-air computation

Fingerprint

Dive into the research topics of 'Over-the-Air Computation Empowered Federated Learning: A Joint Uplink-Downlink Design'. Together they form a unique fingerprint.

Cite this