Generating Differentially Private Datasets Using GANs

Aleksei Triastcyn, Boi Faltings

Abstract: In this paper, we present a technique for generating artificial datasets that retain statistical properties of the real data while providing differential privacy guarantees with respect to this data. We include a Gaussian noise layer in the discriminator of a generative adversarial network to make the output and the gradients differentially private with respect to the training data, and then use the generator component to synthesise privacy-preserving artificial dataset. Our experiments show that under a reasonably small privacy budget we are able to generate data of high quality and successfully train machine learning models on this artificial data.

TL;DR: Train GANs with differential privacy to generate artificial privacy-preserving datasets.

Keywords: generative adversarial networks, differential privacy, synthetic data

Data: [MNIST](https://paperswithcode.com/dataset/mnist), [SVHN](https://paperswithcode.com/dataset/svhn)