Beyond Local Sharpness: Communication-Efficient Global Sharpness-aware Minimization for Federated Learning

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Caldarola, Debora, Cagnasso, Pietro, Caputo, Barbara, Ciccone, Marco
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915218253152256
author Caldarola, Debora
Cagnasso, Pietro
Caputo, Barbara
Ciccone, Marco
author_facet Caldarola, Debora
Cagnasso, Pietro
Caputo, Barbara
Ciccone, Marco
contents Federated learning (FL) enables collaborative model training with privacy preservation. Data heterogeneity across edge devices (clients) can cause models to converge to sharp minima, negatively impacting generalization and robustness. Recent approaches use client-side sharpness-aware minimization (SAM) to encourage flatter minima, but the discrepancy between local and global loss landscapes often undermines their effectiveness, as optimizing for local sharpness does not ensure global flatness. This work introduces FedGloSS (Federated Global Server-side Sharpness), a novel FL approach that prioritizes the optimization of global sharpness on the server, using SAM. To reduce communication overhead, FedGloSS cleverly approximates sharpness using the previous global gradient, eliminating the need for additional client communication. Our extensive evaluations demonstrate that FedGloSS consistently reaches flatter minima and better performance compared to state-of-the-art FL methods across various federated vision benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2412_03752
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Beyond Local Sharpness: Communication-Efficient Global Sharpness-aware Minimization for Federated Learning
Caldarola, Debora
Cagnasso, Pietro
Caputo, Barbara
Ciccone, Marco
Artificial Intelligence
Federated learning (FL) enables collaborative model training with privacy preservation. Data heterogeneity across edge devices (clients) can cause models to converge to sharp minima, negatively impacting generalization and robustness. Recent approaches use client-side sharpness-aware minimization (SAM) to encourage flatter minima, but the discrepancy between local and global loss landscapes often undermines their effectiveness, as optimizing for local sharpness does not ensure global flatness. This work introduces FedGloSS (Federated Global Server-side Sharpness), a novel FL approach that prioritizes the optimization of global sharpness on the server, using SAM. To reduce communication overhead, FedGloSS cleverly approximates sharpness using the previous global gradient, eliminating the need for additional client communication. Our extensive evaluations demonstrate that FedGloSS consistently reaches flatter minima and better performance compared to state-of-the-art FL methods across various federated vision benchmarks.
title Beyond Local Sharpness: Communication-Efficient Global Sharpness-aware Minimization for Federated Learning
topic Artificial Intelligence
url https://arxiv.org/abs/2412.03752