SA-DPSGD: Differentially Private Stochastic Gradient Descent based on Simulated Annealing
share
Differential privacy (DP) provides a formal privacy guarantee that prevents adversaries with access to machine learning models from extracting information about individual training points. Differentially private stochastic gradient descent (DPSGD) is the most popular training method with differential privacy in image recognition. However, existing DPSGD schemes lead to significant performance degradation, which prevents the application of differential privacy. In this paper, we propose a simulated annealing-based differentially private stochastic gradient descent scheme (SA-DPSGD) which accepts a candidate update with a probability that depends both on the update quality and on the number of iterations. Through this random update screening, we make the differentially private gradient descent proceed in the right direction in each iteration, and result in a more accurate model finally. In our experiments, under the same hyperparameters, our scheme achieves test accuracies 98.35 87.41 compared to the state-of-the-art result of 98.12 freely adjusted hyperparameters, our scheme achieves even higher accuracies, 98.89 for closing the accuracy gap between private and non-private image classification.
READ FULL TEXT