Personalized Federated Learning via ADMM with Moreau Envelope

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhu, Shengkun, Zeng, Jinshan, Wang, Sheng, Sun, Yuan, Peng, Zhiyong
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916364732596224
author Zhu, Shengkun
Zeng, Jinshan
Wang, Sheng
Sun, Yuan
Peng, Zhiyong
author_facet Zhu, Shengkun
Zeng, Jinshan
Wang, Sheng
Sun, Yuan
Peng, Zhiyong
contents Personalized federated learning (PFL) is an approach proposed to address the issue of poor convergence on heterogeneous data. However, most existing PFL frameworks require strong assumptions for convergence. In this paper, we propose an alternating direction method of multipliers (ADMM) for training PFL models with Moreau envelope (FLAME), which achieves a sublinear convergence rate, relying on the relatively weak assumption of gradient Lipschitz continuity. Moreover, due to the gradient-free nature of ADMM, FLAME alleviates the need for hyperparameter tuning, particularly in avoiding the adjustment of the learning rate when training the global model. In addition, we propose a biased client selection strategy to expedite the convergence of training of PFL models. Our theoretical analysis establishes the global convergence under both unbiased and biased client selection strategies. Our experiments validate that FLAME, when trained on heterogeneous data, outperforms state-of-the-art methods in terms of model performance. Regarding communication efficiency, it exhibits an average speedup of 3.75x compared to the baselines. Furthermore, experimental results validate that the biased client selection strategy speeds up the convergence of both personalized and global models.
format Preprint
id arxiv_https___arxiv_org_abs_2311_06756
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Personalized Federated Learning via ADMM with Moreau Envelope
Zhu, Shengkun
Zeng, Jinshan
Wang, Sheng
Sun, Yuan
Peng, Zhiyong
Machine Learning
Distributed, Parallel, and Cluster Computing
Personalized federated learning (PFL) is an approach proposed to address the issue of poor convergence on heterogeneous data. However, most existing PFL frameworks require strong assumptions for convergence. In this paper, we propose an alternating direction method of multipliers (ADMM) for training PFL models with Moreau envelope (FLAME), which achieves a sublinear convergence rate, relying on the relatively weak assumption of gradient Lipschitz continuity. Moreover, due to the gradient-free nature of ADMM, FLAME alleviates the need for hyperparameter tuning, particularly in avoiding the adjustment of the learning rate when training the global model. In addition, we propose a biased client selection strategy to expedite the convergence of training of PFL models. Our theoretical analysis establishes the global convergence under both unbiased and biased client selection strategies. Our experiments validate that FLAME, when trained on heterogeneous data, outperforms state-of-the-art methods in terms of model performance. Regarding communication efficiency, it exhibits an average speedup of 3.75x compared to the baselines. Furthermore, experimental results validate that the biased client selection strategy speeds up the convergence of both personalized and global models.
title Personalized Federated Learning via ADMM with Moreau Envelope
topic Machine Learning
Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2311.06756