Hi guys!
I’m approaching differential privacy (github alessio-proietti/spark-mxnet-differential-privacy), I saw there’s a lib for PyTorch which implements DP-SGD. I was wondering if mxnet’s community support something similar.
From what I see at some point in time there was an example in a PR where the authours presumably wanted to publish an attempt to do DP in mxnet. (An example of differentially private deep learning using mxnet gluon by yuxiangw · Pull Request #10779 · apache/incubator-mxnet · GitHub, [WIP] examples for differential privacy by mli · Pull Request #14707 · apache/incubator-mxnet · GitHub)
What is the state of the affair? Is DP in mxnet a thing, someone is reasearching on it?
Thanks for listening,
Alessio