Differentially private synthetic medical data generation using convolutional GANs

Deep learning models have demonstrated superior performance in several real-world application problems such as image classification and speech processing. However, creating these sensitive domains like healthcare typically requires addressing certain privacy challenges that bring unique concerns. One effective way to handle private data concerns is generate realistic synthetic can provide practically acceptable quality well be used improve model performance. To tackle this challenge, we develop a differentially framework for generation using Rényi differential . Our approach builds on convolutional autoencoders generative adversarial networks preserve critical characteristics of the generated data. In addition, our capture temporal information feature correlations present original We demonstrate outperforms existing state-of-the-art under same budget publicly available benchmark medical datasets both supervised unsupervised settings. The source code work at https://github.com/astorfi/differentially-private-cgan.