Federated K-means Clustering

TL;DR

This paper introduces a federated K-means clustering algorithm that enables unsupervised learning across distributed datasets while addressing challenges like varying cluster numbers and convergence issues on less separable data.

Contribution

It presents a novel federated K-means algorithm that handles varying cluster counts and improves convergence on challenging datasets, filling a gap in unsupervised federated learning.

Findings

Effective clustering on distributed data without data pooling

Addresses varying number of clusters in federated settings

Improves convergence on less separable datasets

Abstract

Federated learning is a technique that enables the use of distributed datasets for machine learning purposes without requiring data to be pooled, thereby better preserving privacy and ownership of the data. While supervised FL research has grown substantially over the last years, unsupervised FL methods remain scarce. This work introduces an algorithm which implements K-means clustering in a federated manner, addressing the challenges of varying number of clusters between centers, as well as convergence on less separable datasets.