FedTune: Automatic Tuning of Federated Learning Hyper-Parameters from System Perspective

Huanle Zhang,Xin Liu,Mi Zhang,Prasant Mohapatra,Michael Delucia

doi:10.1109/milcom55135.2022.10017717

FedTune: Automatic Tuning of Federated Learning Hyper-Parameters from System Perspective

Huanle Zhang, Xin Liu + Show 3 more

Open Access

https://doi.org/10.1109/milcom55135.2022.10017717

Copy DOI

Publication Date: Nov 28, 2022

Citations: 5

Affiliation: University of California, Davis, The Ohio State University, DEVCOM Army Research Laboratory

#Federated Learning #Automatic Tuning + Show 8 more

Abstract
Full-Text PDF
Similar Papers

Abstract

In Federated Learning (FL), hyper-parameters significantly affect the training overhead in terms of computation time, transmission time, computation load, and transmission load. The current practice of manually selecting FL hyper-parameters puts a high burden on FL practitioners since various applications have different training preferences. In this paper, we propose FedTune, an automatic hyper-parameter tuning algorithm tailored to applications' diverse system requirements in FL training. FedTune is lightweight and flexible, achieving 8.48%-26.75% improvement for different datasets compared to using fixed FL hyper-parameters.

Full Text