An Adaptive Differential Privacy Method Based on Federated Learning

TL;DR

This paper introduces an adaptive differential privacy approach for federated learning that dynamically adjusts privacy budgets based on multiple training factors, improving privacy-accuracy balance.

Contribution

It proposes a novel adaptive privacy mechanism considering various influencing factors, enhancing privacy protection without sacrificing model accuracy.

Findings

Reduces privacy budget by about 16%

Maintains comparable training accuracy

Analyzes parameter ranges and privacy guarantees

Abstract

Differential privacy is one of the methods to solve the problem of privacy protection in federated learning. Setting the same privacy budget for each round will result in reduced accuracy in training. The existing methods of the adjustment of privacy budget consider fewer influencing factors and tend to ignore the boundaries, resulting in unreasonable privacy budgets. Therefore, we proposed an adaptive differential privacy method based on federated learning. The method sets the adjustment coefficient and scoring function according to accuracy, loss, training rounds, and the number of datasets and clients. And the privacy budget is adjusted based on them. Then the local model update is processed according to the scaling factor and the noise. Fi-nally, the server aggregates the noised local model update and distributes the noised global model. The range of parameters and the privacy of the method are analyzed. Through the experimental evaluation, it can reduce the privacy budget by about 16%, while the accuracy remains roughly the same.