Efficient and Light-Weight Federated Learning via Asynchronous Distributed Dropout
Chen Dun, Mirian Hipolito, Chris Jermaine, Dimitrios Dimitriadis, Anastasios Kyrillidis
Abstract
Asynchronous learning protocols have regained attention lately, especially in the Federated Learning (FL) setup, where slower clients can severely impede the learning process. Herein, we propose AsyncDrop, a novel asynchronous FL framework that utilizes dropout regularization to handle device heterogeneity in distributed settings. Overall, AsyncDrop achieves better performance compared to state of the art asynchronous methodologies, while resulting in less communication and training time overheads. The key idea revolves around creating “submodels” out of the global model, and distributing their training to workers, based on device heterogeneity. We rigorously justify that such an approach can be theoretically characterized. We implement our approach and compare it against other asynchronous baselines, both by design and by adapting existing synchronous FL algorithms to asynchronous scenarios. Empirically, AsyncDrop reduces the communication cost and training time, while matching or improving the final test accuracy in diverse non-i.i.d. FL scenarios.
BibTeX
@InProceedings{pmlr-v206-dun23a,
title = {Efficient and Light-Weight Federated Learning via Asynchronous Distributed Dropout},
author = {Dun, Chen and Hipolito, Mirian and Jermaine, Chris and Dimitriadis, Dimitrios and Kyrillidis, Anastasios},
booktitle = {Proceedings of The 26th International Conference on Artificial Intelligence and Statistics},
pages = {6630--6660},
year = {2023},
editor = {Ruiz, Francisco and Dy, Jennifer and van de Meent, Jan-Willem},
volume = {206},
series = {Proceedings of Machine Learning Research},
month = {25--27 Apr},
publisher = {PMLR},
pdf = {https://proceedings.mlr.press/v206/dun23a/dun23a.pdf},
url = {https://proceedings.mlr.press/v206/dun23a.html},
abstract = {Asynchronous learning protocols have regained attention lately, especially in the Federated Learning (FL) setup, where slower clients can severely impede the learning process. Herein, we propose AsyncDrop, a novel asynchronous FL framework that utilizes dropout regularization to handle device heterogeneity in distributed settings. Overall, AsyncDrop achieves better performance compared to state of the art asynchronous methodologies, while resulting in less communication and training time overheads. The key idea revolves around creating “submodels” out of the global model, and distributing their training to workers, based on device heterogeneity. We rigorously justify that such an approach can be theoretically characterized. We implement our approach and compare it against other asynchronous baselines, both by design and by adapting existing synchronous FL algorithms to asynchronous scenarios. Empirically, AsyncDrop reduces the communication cost and training time, while matching or improving the final test accuracy in diverse non-i.i.d. FL scenarios.}
}