Communication-Efficient Stochastic Zeroth-Order Optimization for Federated Learning
Abstract
Federated learning (FL), as an emerging edge artificial intelligence paradigm, enables many edge devices to collaboratively train a global model without sharing their private data. To enhance the training efficiency of FL, various algorithms have been proposed, ranging from first-order to second-order methods. However, these algorithms cannot be applied in scenarios where the gradient information is not available, e.g., federated black-box attack and federated hyperparameter tuning. To address this issue, in this paper we propose a derivative-free federated zeroth-order optimization (FedZO) algorithm featured by performing multiple local updates based on stochastic gradient estimators in each communication round and enabling partial device participation. Under non-convex settings, we derive the convergence performance of the FedZO algorithm on non-independent and identically distributed data and characterize the impact of the numbers of local iterates and participating edge devices on the convergence. To enable communication-efficient FedZO over wireless networks, we further propose an over-the-air computation (AirComp) assisted FedZO algorithm. With an appropriate transceiver design, we show that the convergence of AirComp-assisted FedZO can still be preserved under certain signal-to-noise ratio conditions. Simulation results demonstrate the effectiveness of the FedZO algorithm and validate the theoretical observations.
Cite
@article{arxiv.2201.09531,
title = {Communication-Efficient Stochastic Zeroth-Order Optimization for Federated Learning},
author = {Wenzhi Fang and Ziyi Yu and Yuning Jiang and Yuanming Shi and Colin N. Jones and Yong Zhou},
journal= {arXiv preprint arXiv:2201.09531},
year = {2022}
}
Comments
This work was accepted to Transaction on Signal Processing