An Asynchronous Federated Learning Aggregation Method Based on Adaptive Differential Privacy

Federated learning is a distributed machine learning technique that enables multiple devices to collaborate on learning a shared model without exchanging data. It can be used to improve model accuracy while protecting user privacy. However, traditional federated learning is vulnerable to attacks from generative adversarial networks (GANs). As a new privacy protection method, differential privacy enhances privacy protection capabilities by sacrificing some data accuracy. We optimized the privacy budget allocation scheme in traditional differential privacy and proposed an adaptive parameter-based differential privacy method that improves training accuracy while maintaining the overall privacy budget. Additionally, we proposed an asynchronous federated learning aggregation scheme that combines privacy budget and freshness, reducing the impact of differential privacy on accuracy. We conducted extensive experiments on Gaussian mechanism-based differential privacy and Laplace mechanism-based differential privacy algorithms. Experimental results show that, under the same privacy budget, our algorithm achieves higher accuracy and lower communication overhead compared to baseline algorithms.