Pinching Antennas-Assisted Low-Latency Federated Learning Over Multi-User Wireless Networks

TL;DR

This paper introduces FedPASS, a framework that leverages pinching-antenna systems to dynamically improve wireless communication for federated learning, significantly reducing training latency while maintaining accuracy.

Contribution

It develops a novel optimization framework and algorithms for low-latency federated learning using PASS technology, addressing practical constraints and improving communication reliability.

Findings

FedPASS achieves comparable accuracy to ideal baselines.

It drastically reduces total training latency.

Numerical results on MNIST and CIFAR-10 validate effectiveness.

Abstract

Federated learning (FL) over wireless networks is fundamentally constrained by unreliable communication links, particularly when uplink channels suffer from blockage, fading, or weak line-of-sight (LoS) conditions. Pinching-antenna systems (PASSs) offer a new physical-layer capability to dynamically reposition radiating points along a dielectric waveguide, enabling controllable LoS connectivity and significantly improved channel quality. This paper develops FedPASS, a novel framework for low-latency wireless FL assisted by PASS. We formulate a multi-objective optimization problem that jointly minimizes the end-to-end round latency and an upper bound on the FL optimality gap. The resulting formulation is a mixed-integer nonlinear program subject to practical constraints on scheduling, transmit power, local CPU frequency, and PA placement. To address the resulting computational challenges, we develop a two-tier iterative algorithm: an outer loop that updates scheduling, communication time allocation, and power control via block coordinate descent, and an inner loop that optimizes PA locations using a Gauss-Seidel-based coordinate update with grid search under spacing constraints. Numerical results on MNIST and CIFAR-10 demonstrate that FedPASS achieves accuracy comparable to idealized FL baselines while drastically reducing the total training latency compared to conventional wireless FL.