Fetching the paper…

Improving Deep Learning with Differential Privacy using Gradient Encoding and Denoising · Around