Joint Downlink-Uplink Beamforming for Wireless Multi-Antenna Federated Learning

TL;DR

This paper proposes a joint downlink-uplink beamforming method for wireless federated learning with multi-antenna base stations, improving communication efficiency and model accuracy over noisy channels.

Contribution

It introduces a novel low-complexity joint beamforming algorithm that optimizes both downlink and uplink transmissions in wireless FL systems.

Findings

Significantly outperforms separate-link design approaches

Nearly matches ideal FL performance with error-free links

Provides a practical solution for wireless FL communication challenges

Abstract

We study joint downlink-uplink beamforming design for wireless federated learning (FL) with a multi-antenna base station. Considering analog transmission over noisy channels and uplink over-the-air aggregation, we derive the global model update expression over communication rounds. We then obtain an upper bound on the expected global loss function, capturing the downlink and uplink beamforming and receiver noise effect. We propose a low-complexity joint beamforming algorithm to minimize this upper bound, which employs alternating optimization to breakdown the problem into three subproblems, each solved via closed-form gradient updates. Simulation under practical wireless system setup shows that our proposed joint beamforming design solution substantially outperforms the conventional separate-link design approach and nearly attains the performance of ideal FL with error-free communication links.