On-Demand Federated Learning for Arbitrary Target Class Distributions

Isu Jeong, Seulki Lee
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, PMLR 238:3421-3429, 2024.

Abstract

We introduce On-Demand Federated Learning (On-Demand FL), which enables on-demand federated learning of a deep model for an arbitrary target data distribution of interest by making the best use of the heterogeneity (non-IID-ness) of local client data, unlike existing approaches trying to circumvent the non-IID nature of federated learning. On-Demand FL composes a dataset of the target distribution, which we call the composite dataset, from a selected subset of local clients whose aggregate distribution is expected to emulate the target distribution as a whole. As the composite dataset consists of a precise yet diverse subset of clients reflecting the target distribution, the on-demand model trained with exactly enough selected clients becomes able to improve the model performance on the target distribution compared when trained with off-target and/or unknown distributions while reducing the number of participating clients and federating rounds. We model the target data distribution in terms of class and estimate the class distribution of each local client from the weight gradient of its local model. Our experiment results show that On-Demand FL achieves up to 5% higher classification accuracy on various target distributions just involving 9${\times}$ fewer clients with FashionMNIST, CIFAR-10, and CIFAR-100.

Cite this Paper

BibTeX
@InProceedings{pmlr-v238-jeong24a, title = {On-Demand Federated Learning for Arbitrary Target Class Distributions}, author = {Jeong, Isu and Lee, Seulki}, booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, pages = {3421--3429}, year = {2024}, editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, volume = {238}, series = {Proceedings of Machine Learning Research}, month = {02--04 May}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v238/jeong24a/jeong24a.pdf}, url = {https://proceedings.mlr.press/v238/jeong24a.html}, abstract = {We introduce On-Demand Federated Learning (On-Demand FL), which enables on-demand federated learning of a deep model for an arbitrary target data distribution of interest by making the best use of the heterogeneity (non-IID-ness) of local client data, unlike existing approaches trying to circumvent the non-IID nature of federated learning. On-Demand FL composes a dataset of the target distribution, which we call the composite dataset, from a selected subset of local clients whose aggregate distribution is expected to emulate the target distribution as a whole. As the composite dataset consists of a precise yet diverse subset of clients reflecting the target distribution, the on-demand model trained with exactly enough selected clients becomes able to improve the model performance on the target distribution compared when trained with off-target and/or unknown distributions while reducing the number of participating clients and federating rounds. We model the target data distribution in terms of class and estimate the class distribution of each local client from the weight gradient of its local model. Our experiment results show that On-Demand FL achieves up to 5% higher classification accuracy on various target distributions just involving 9${\times}$ fewer clients with FashionMNIST, CIFAR-10, and CIFAR-100.} }
Endnote
%0 Conference Paper %T On-Demand Federated Learning for Arbitrary Target Class Distributions %A Isu Jeong %A Seulki Lee %B Proceedings of The 27th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2024 %E Sanjoy Dasgupta %E Stephan Mandt %E Yingzhen Li %F pmlr-v238-jeong24a %I PMLR %P 3421--3429 %U https://proceedings.mlr.press/v238/jeong24a.html %V 238 %X We introduce On-Demand Federated Learning (On-Demand FL), which enables on-demand federated learning of a deep model for an arbitrary target data distribution of interest by making the best use of the heterogeneity (non-IID-ness) of local client data, unlike existing approaches trying to circumvent the non-IID nature of federated learning. On-Demand FL composes a dataset of the target distribution, which we call the composite dataset, from a selected subset of local clients whose aggregate distribution is expected to emulate the target distribution as a whole. As the composite dataset consists of a precise yet diverse subset of clients reflecting the target distribution, the on-demand model trained with exactly enough selected clients becomes able to improve the model performance on the target distribution compared when trained with off-target and/or unknown distributions while reducing the number of participating clients and federating rounds. We model the target data distribution in terms of class and estimate the class distribution of each local client from the weight gradient of its local model. Our experiment results show that On-Demand FL achieves up to 5% higher classification accuracy on various target distributions just involving 9${\times}$ fewer clients with FashionMNIST, CIFAR-10, and CIFAR-100.
APA
Jeong, I. & Lee, S.. (2024). On-Demand Federated Learning for Arbitrary Target Class Distributions. Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 238:3421-3429 Available from https://proceedings.mlr.press/v238/jeong24a.html.

Related Material