Asynchronous Federated Learning with Differential Privacy for Edge Intelligence

TL;DR

This paper proposes a differentially private asynchronous federated learning framework for edge-cloud systems, introducing a multi-stage adjustable algorithm that enhances model accuracy and convergence while maintaining privacy.

Contribution

It provides the first formal analysis of model convergence in asynchronous federated learning under differential privacy and introduces MAPA, a dynamic algorithm to optimize privacy-utility trade-offs.

Findings

MAPA improves model accuracy and convergence speed.

The approach maintains strong privacy guarantees.

Extensive experiments validate the effectiveness of MAPA.

Abstract

Federated learning has been showing as a promising approach in paving the last mile of artificial intelligence, due to its great potential of solving the data isolation problem in large scale machine learning. Particularly, with consideration of the heterogeneity in practical edge computing systems, asynchronous edge-cloud collaboration based federated learning can further improve the learning efficiency by significantly reducing the straggler effect. Despite no raw data sharing, the open architecture and extensive collaborations of asynchronous federated learning (AFL) still give some malicious participants great opportunities to infer other parties' training data, thus leading to serious concerns of privacy. To achieve a rigorous privacy guarantee with high utility, we investigate to secure asynchronous edge-cloud collaborative federated learning with differential privacy, focusing on the impacts of differential privacy on model convergence of AFL. Formally, we give the first analysis on the model convergence of AFL under DP and propose a multi-stage adjustable private algorithm (MAPA) to improve the trade-off between model utility and privacy by dynamically adjusting both the noise scale and the learning rate. Through extensive simulations and real-world experiments with an edge-could testbed, we demonstrate that MAPA significantly improves both the model accuracy and convergence speed with sufficient privacy guarantee.