Abstract:Federated learning (FL) is an emerging distributed machine learning paradigm that enables collaborative training of machine learning models over decentralized devices without exposing their local data. One of the major challenges in FL is the presence of uneven data distributions across client devices, violating the well-known assumption of independent-and-identically-distributed (IID) training samples in conventional machine learning. To address the performance degradation issue incurred by such data heterogeneity, clustered federated learning (CFL) shows its promise by grouping clients into separate learning clusters based on the similarity of their local data distributions. However, state-of-the-art CFL approaches require a large number of communication rounds to learn the distribution similarities during training until the formation of clusters is stabilized. Moreover, some of these algorithms heavily rely on a predefined number of clusters, thus limiting their flexibility and adaptability. In this paper, we propose {\em FedClust}, a novel approach for CFL that leverages the correlation between local model weights and the data distribution of clients. {\em FedClust} groups clients into clusters in a one-shot manner by measuring the similarity degrees among clients based on the strategically selected partial weights of locally trained models. We conduct extensive experiments on four benchmark datasets with different non-IID data settings. Experimental results demonstrate that {\em FedClust} achieves higher model accuracy up to $\sim$45\% as well as faster convergence with a significantly reduced communication cost up to 2.7$\times$ compared to its state-of-the-art counterparts.

Comparative Evaluation of Clustered Federated Learning Methods

Clustered Federated Learning in Heterogeneous Environment.

FedClust: Tackling Data Heterogeneity in Federated Learning through Weight-Driven Client Clustering

Fuzzy Clustered Federated Learning under Mixed Data Distributions.

Flexible Clustered Federated Learning for Client-Level Data Distribution Shift

Genetic CFL: Hyperparameter Optimization in Clustered Federated Learning

Genetic CFL: Optimization of Hyper-Parameters in Clustered Federated Learning

Heterogeneity-aware device selection for clustered federated learning in IoT

Contrastive encoder pre-training-based clustered federated learning for heterogeneous data

A Comprehensive Study on Personalized Federated Learning with Non-IID Data.

FedGroup: Efficient Clustered Federated Learning via Decomposed Data-Driven Measure

FedAC: An Adaptive Clustered Federated Learning Framework for Heterogeneous Data

A Generalized Look at Federated Learning: Survey and Perspectives

Towards Hierarchical Clustered Federated Learning with Model Stability on Mobile Devices

Optimisation of federated learning settings under statistical heterogeneity variations

FedClust: Optimizing Federated Learning on Non-IID Data through Weight-Driven Client Clustering

A Survey of Federated Learning on Non-IID Data

Stochastic Clustered Federated Learning

On the Impact of Data Heterogeneity in Federated Learning Environments with Application to Healthcare Networks

Communication-efficient clustered federated learning via model distance

Completely Heterogeneous Federated Learning