Advanced ML encryption techniques in practice

In addition to SEAL, other libraries support advanced encryption and privacy-preserving techniques. This includes OpenMined PySyft, which supports encrypted computation, FL, and DP. You can find more about PySyft at https://github.com/OpenMined/PySyft.

Additionally, tf-encrypted is a framework for encrypted deep learning in TensorFlow 2.x. The genomic study we discussed earlier uses tf-secure. The focus of tf-encrypted is on secure MPC and encrypting gradients and parameters so that they can safely collaborate without revealing input data.

The following is an example of using tf-encrypted with Keras to implement secure computation by encrypting the gradients and other model parameters in a Keras DNN. The example is taken from the tf-encrypted home page at https://tf-encrypted.io/, using network building blocks from tf-encrypted:

import tensorflow as tf
import tf_encrypted as tfe
@tfe.local_computation('prediction-client&apos...