Enhanced Over-the-Air Federated Learning Using AI-based Fluid Antenna System

TL;DR

This paper proposes an AI-based fluid antenna system for over-the-air federated learning, improving convergence and performance by optimizing antenna positions and transmission parameters through advanced algorithms.

Contribution

It introduces a fluid antenna approach for OTA-FL, formulates a joint optimization problem, and develops a reinforcement learning algorithm to enhance learning efficiency.

Findings

Fluid antennas improve OTA-FL convergence.

Optimized antenna positioning enhances learning performance.

Proposed RDPG algorithm outperforms existing methods.

Abstract

This paper investigates an over-the-air federated learning (OTA-FL) system that employs fluid antennas (FAs) at an access point. The system enhances learning performance by leveraging the additional degrees of freedom provided by antenna mobility. We analyze the convergence of the OTA-FL system and derive the optimality gap to illustrate the influence of FAs on learning performance. With these results, we formulate a nonconvex optimization problem to minimize the optimality gap by jointly optimizing the positions of the FAs, the beamforming vector, and the transmit power allocation at each user. To address the dynamic environment, we cast this optimization problem as a Markov decision process and propose the recurrent deterministic policy gradient (RDPG) algorithm. Finally, extensive simulations show that the FA-assisted OTA-FL system outperforms systems with fixed-position antennas and that the RDPG algorithm surpasses the existing methods.