Communication-Efficient Stochastic Zeroth-Order Optimization for Federated Learning

TL;DR

This paper introduces a communication-efficient zeroth-order federated learning algorithm that works without gradient information, suitable for black-box scenarios, and enhances efficiency through local updates and over-the-air computation.

Contribution

It proposes the FedZO algorithm for derivative-free federated learning, including an AirComp-assisted version for wireless networks, with theoretical convergence analysis and practical validation.

Findings

FedZO converges under non-convex, non-i.i.d. data.

Over-the-air FedZO maintains convergence with proper transceiver design.

Simulation confirms effectiveness and theoretical insights.

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.