# SA-DPSGD: Differentially Private Stochastic Gradient Descent based on Simulated Annealing Source: [SA-DPSGD: Differentially Private Stochastic Gradient Descent based on Simulated Annealing]() Collection index: [llms.txt]() Data updated: 2025-10-26T00:51:11.369Z Authors: Jie Fu, Zhili Chen, XinPeng Ling Published: 2022-11-14T09:20:48.000Z ## Abstract 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% and 60.92% on datasets MNIST, FashionMNIST and CIFAR10, respectively, compared to the state-of-the-art result of 98.12%, 86.33% and 59.34%. Under the freely adjusted hyperparameters, our scheme achieves even higher accuracies, 98.89%, 88.50% and 64.17%. We believe that our method has a great contribution for closing the accuracy gap between private and non-private image classification. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Classification on MNIST (test) (Epsilon-based Accuracy)](): Image Classification; MNIST (test); Accuracy (\\epsilon=1): 93.01. - [Image Classification on FMNIST (test) (Epsilon-Weighted Accuracy)](): Image Classification; fMNIST (test); Accuracy (epsilon=1): 82.08. ## Original sources and implementation [Original paper]() [PDF]()