DRACO: Decentralized Asynchronous Federated Learning over Row-Stochastic Wireless Networks

TL;DR

DRACO introduces a decentralized asynchronous SGD method for wireless IoT and Edge AI networks, enabling stable convergence without synchronization, thus improving autonomous training in fully decentralized environments.

Contribution

The paper presents a novel asynchronous decentralized SGD algorithm over row-stochastic wireless networks, allowing continuous communication and decoupled schedules for improved stability and autonomy.

Findings

Convergence analysis shows stability without synchronization.

Numerical experiments validate effectiveness in decentralized settings.

Decoupled communication and computation schedules enhance autonomy.

Abstract

Recent developments and emerging use cases, such as smart Internet of Things (IoT) and Edge AI, have sparked considerable interest in the training of neural networks over fully decentralized (serverless) networks. One of the major challenges of decentralized learning is to ensure stable convergence without resorting to strong assumptions applied for each agent regarding data distributions or updating policies. To address these issues, we propose DRACO, a novel method for decentralized asynchronous Stochastic Gradient Descent (SGD) over row-stochastic gossip wireless networks by leveraging continuous communication. Our approach enables edge devices within decentralized networks to perform local training and model exchanging along a continuous timeline, thereby eliminating the necessity for synchronized timing. The algorithm also features a specific technique of decoupling communication and computation schedules, which empowers complete autonomy for all users and manageable instructions for stragglers. Through a comprehensive convergence analysis, we highlight the advantages of asynchronous and autonomous participation in decentralized optimization. Our numerical experiments corroborate the efficacy of the proposed technique.