Heterogeneity for the Win: One-Shot Federated Clustering
Don Kurian Dennis, Tian Li, Virginia Smith
Abstract
In this work, we explore the unique challenges—and opportunities—of unsupervised federated learning (FL). We develop and analyze a one-shot federated clustering scheme, kfed, based on the widely-used Lloyd’s method for $k$-means clustering. In contrast to many supervised problems, we show that the issue of statistical heterogeneity in federated networks can in fact benefit our analysis. We analyse kfed under a center separation assumption and compare it to the best known requirements of its centralized counterpart. Our analysis shows that in heterogeneous regimes where the number of clusters per device $(k’)$ is smaller than the total number of clusters over the network $k$, $(k’\le \sqrt{k})$, we can use heterogeneity to our advantage—significantly weakening the cluster separation requirements for kfed. From a practical viewpoint, kfed also has many desirable properties: it requires only round of communication, can run asynchronously, and can handle partial participation or node/network failures. We motivate our analysis with experiments on common FL benchmarks, and highlight the practical utility of one-shot clustering through use-cases in personalized FL and device sampling.
BibTeX
@InProceedings{pmlr-v139-dennis21a,
title = {Heterogeneity for the Win: One-Shot Federated Clustering},
author = {Dennis, Don Kurian and Li, Tian and Smith, Virginia},
booktitle = {Proceedings of the 38th International Conference on Machine Learning},
pages = {2611--2620},
year = {2021},
editor = {Meila, Marina and Zhang, Tong},
volume = {139},
series = {Proceedings of Machine Learning Research},
month = {18--24 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v139/dennis21a/dennis21a.pdf},
url = {https://proceedings.mlr.press/v139/dennis21a.html},
abstract = {In this work, we explore the unique challenges—and opportunities—of unsupervised federated learning (FL). We develop and analyze a one-shot federated clustering scheme, kfed, based on the widely-used Lloyd’s method for $k$-means clustering. In contrast to many supervised problems, we show that the issue of statistical heterogeneity in federated networks can in fact benefit our analysis. We analyse kfed under a center separation assumption and compare it to the best known requirements of its centralized counterpart. Our analysis shows that in heterogeneous regimes where the number of clusters per device $(k’)$ is smaller than the total number of clusters over the network $k$, $(k’\le \sqrt{k})$, we can use heterogeneity to our advantage—significantly weakening the cluster separation requirements for kfed. From a practical viewpoint, kfed also has many desirable properties: it requires only round of communication, can run asynchronously, and can handle partial participation or node/network failures. We motivate our analysis with experiments on common FL benchmarks, and highlight the practical utility of one-shot clustering through use-cases in personalized FL and device sampling.}
}