Over-the-Air Federated Learning via Weighted Aggregation

TL;DR

This paper proposes an over-the-air federated learning scheme with adaptive weighted aggregation that improves learning accuracy by mitigating wireless channel effects without requiring channel state information at transmitters.

Contribution

It introduces a novel adaptive weighting method for over-the-air federated learning that enhances robustness against wireless channel variability and device heterogeneity.

Findings

Achieves 15% higher accuracy than schemes with CSIT

Surpasses non-CSIT schemes by 30% in accuracy

Provides convergence analysis and optimized weight algorithms

Abstract

This paper introduces a new federated learning scheme that leverages over-the-air computation. A novel feature of this scheme is the proposal to employ adaptive weights during aggregation, a facet treated as predefined in other over-the-air schemes. This can mitigate the impact of wireless channel conditions on learning performance, without needing channel state information at transmitter side (CSIT). We provide a mathematical methodology to derive the convergence bound for the proposed scheme in the context of computational heterogeneity and general loss functions, supplemented with design insights. Accordingly, we propose aggregation cost metrics and efficient algorithms to find optimized weights for the aggregation. Finally, through numerical experiments, we validate the effectiveness of the proposed scheme. Even with the challenges posed by channel conditions and device heterogeneity, the proposed scheme surpasses other over-the-air strategies by an accuracy improvement of 15% over the scheme using CSIT and 30% compared to the one without CSIT.