arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2601.04483cs.LGcs.AIcs.ITeess.SPmath.IT
收藏

混合联邦学习用于抗噪训练

Hybrid Federated Learning for Noise-Robust Training

School of Electrical Engineering, KAIST(韩国科学技术院电子工程学院);School of Electronics Engineering, Kyungpook National University(庆北国立大学电子工程学院)

Yongjun Kim, Hyeongjun Park, Hwanjin Kim, Junil Choi

AI总结:

本文提出混合联邦学习框架,通过自适应聚类和权重选择提升抗噪训练性能,在低信噪比下实现更高的测试准确率。

摘要:

联邦学习(FL)和联邦蒸馏(FD)是分布式学习范式,通过增强隐私训练UE模型,各自在噪声鲁棒性和学习速度之间提供不同的权衡。为缓解各自缺点,我们提出一种混合联邦学习(HFL)框架,其中每个用户设备(UE)传输梯度或logits,基站(BS)选择FL和FD更新的每轮权重。我们推导了HFL框架的收敛性,并引入两种方法利用HFL中的自由度(DoF),即(i)通过Jenks优化进行自适应UE聚类,(ii)通过阻尼牛顿法进行自适应权重选择。数值结果表明,当同时利用两种DoF时,HFL在低SNR下实现更优的测试准确率。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑