Quantized Analog Beamforming Enabled Multi-task Federated Learning Over-the-air

2025-03-22

AI Summary

To address severe inter-task interference in over-the-air computation (AirComp)-enabled multi-task federated learning (FL) over resource-constrained wireless networks, this paper proposes a receiver-side quantized analog beamforming scheme. It is the first to deeply integrate statistical interference cancellation with low-precision quantized beamforming. We theoretically prove that the residual interference power converges at rate $O(1/N_r)$, independent of phase shifter quantization resolution. Leveraging channel hardening and favorable propagation in massive MIMO, we derive closed-form beamformers. Numerical results demonstrate that introducing only a small number of low-resolution phase shifters enables near-optimal FL performance—achieving convergence accuracy close to the ideal upper bound—while simultaneously enhancing communication efficiency and hardware feasibility.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Over-the-air computation (AirComp) has recently emerged as a pivotal technique for communication-efficient federated learning (FL) in resource-constrained wireless networks. Though AirComp leverages the superposition property of multiple access channels for computation, it inherently limits its ability to manage inter-task interference in multi-task computing. In this paper, we propose a quantized analog beamforming scheme at the receiver to enable simultaneous multi-task FL. Specifically, inspiring by the favorable propagation and channel hardening properties of large-scale antenna arrays, a targeted analog beamforming method in closed form is proposed for statistical interference elimination. Analytical results reveal that the interference power vanishes by an order of $mathcal{O}left(1/N_r ight)$ with the number of analog phase shifters, $N_r$, irrespective of their quantization precision. Numerical results demonstrate the effectiveness of the proposed analog beamforming method and show that the performance upper bound of ideal learning without errors can be achieved by increasing the number of low-precision analog phase shifters.

Problem

• Enables multi-task federated learning via quantized analog beamforming
• Reduces inter-task interference in wireless AirComp systems
• Achieves ideal learning performance with low-precision phase shifters

Innovation

• Quantized analog beamforming for multi-task FL
• Large-scale antenna arrays eliminate interference
• Low-precision phase shifters achieve ideal performance