Hybrid Federated Learning for Noise-Robust Training

TL;DR

This paper introduces a hybrid federated learning framework that combines federated learning and federated distillation to improve noise robustness and training efficiency in distributed models, with adaptive methods for optimization.

Contribution

The paper proposes a novel hybrid federated learning framework with adaptive strategies for weight selection and clustering, enhancing noise robustness and convergence.

Findings

HFL achieves higher test accuracy at low SNR.

Adaptive DoF exploitation improves learning performance.

Convergence of the HFL framework is theoretically derived.

Abstract

Federated learning (FL) and federated distillation (FD) are distributed learning paradigms that train UE models with enhanced privacy, each offering different trade-offs between noise robustness and learning speed. To mitigate their respective weaknesses, we propose a hybrid federated learning (HFL) framework in which each user equipment (UE) transmits either gradients or logits, and the base station (BS) selects the per-round weights of FL and FD updates. We derive convergence of HFL framework and introduce two methods to exploit degrees of freedom (DoF) in HFL, which are (i) adaptive UE clustering via Jenks optimization and (ii) adaptive weight selection via a damped Newton method. Numerical results show that HFL achieves superior test accuracy at low SNR when both DoF are exploited.