Hybrid Federated Learning for Noise-Robust Training

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Kim, Yongjun, Park, Hyeongjun, Kim, Hwanjin, Choi, Junil
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908753034477568
author Kim, Yongjun
Park, Hyeongjun
Kim, Hwanjin
Choi, Junil
author_facet Kim, Yongjun
Park, Hyeongjun
Kim, Hwanjin
Choi, Junil
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2601_04483
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Hybrid Federated Learning for Noise-Robust Training
Kim, Yongjun
Park, Hyeongjun
Kim, Hwanjin
Choi, Junil
Machine Learning
Artificial Intelligence
Information Theory
Signal Processing
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.
title Hybrid Federated Learning for Noise-Robust Training
topic Machine Learning
Artificial Intelligence
Information Theory
Signal Processing
url https://arxiv.org/abs/2601.04483