Adaptive Guidance for Local Training in Heterogeneous Federated Learning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhang, Jianqing, Liu, Yang, Hua, Yang, Cao, Jian, Yang, Qiang
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916773395169280
author Zhang, Jianqing
Liu, Yang
Hua, Yang
Cao, Jian
Yang, Qiang
author_facet Zhang, Jianqing
Liu, Yang
Hua, Yang
Cao, Jian
Yang, Qiang
contents Model heterogeneity poses a significant challenge in Heterogeneous Federated Learning (HtFL). In scenarios with diverse model architectures, directly aggregating model parameters is impractical, leading HtFL methods to incorporate an extra objective alongside the original local objective on each client to facilitate collaboration. However, this often results in a mismatch between the extra and local objectives. To resolve this, we propose Federated Learning-to-Guide (FedL2G), a method that adaptively learns to guide local training in a federated manner, ensuring the added objective aligns with each client's original goal. With theoretical guarantees, FedL2G utilizes only first-order derivatives w.r.t. model parameters, achieving a non-convex convergence rate of O(1/T). We conduct extensive experiments across two data heterogeneity and six model heterogeneity settings, using 14 heterogeneous model architectures (e.g., CNNs and ViTs). The results show that FedL2G significantly outperforms seven state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2410_06490
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Adaptive Guidance for Local Training in Heterogeneous Federated Learning
Zhang, Jianqing
Liu, Yang
Hua, Yang
Cao, Jian
Yang, Qiang
Machine Learning
Artificial Intelligence
Model heterogeneity poses a significant challenge in Heterogeneous Federated Learning (HtFL). In scenarios with diverse model architectures, directly aggregating model parameters is impractical, leading HtFL methods to incorporate an extra objective alongside the original local objective on each client to facilitate collaboration. However, this often results in a mismatch between the extra and local objectives. To resolve this, we propose Federated Learning-to-Guide (FedL2G), a method that adaptively learns to guide local training in a federated manner, ensuring the added objective aligns with each client's original goal. With theoretical guarantees, FedL2G utilizes only first-order derivatives w.r.t. model parameters, achieving a non-convex convergence rate of O(1/T). We conduct extensive experiments across two data heterogeneity and six model heterogeneity settings, using 14 heterogeneous model architectures (e.g., CNNs and ViTs). The results show that FedL2G significantly outperforms seven state-of-the-art methods.
title Adaptive Guidance for Local Training in Heterogeneous Federated Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2410.06490