A privacy-preserving protocol for neural-network-based computation
Mauro Barni, Claudio Orlandi, Alessandro Piva · 2006
The problem of secure data processing by means of a neural network (NN) is addressed. Secure processing refers to the possibility that the NN owner does not get any knowledge about the processed data since they are provided to him in encrypted format. At the same time, the NN itself is protected, given that its owner may not be willing to disclose the knowledge embedded within it. Two different levels of protection are considered: according to the first one only the NN weights are protected, whereas the second level also permits to protect the node activation functions. An efficient way of implementing the proposed protocol by means of some recently proposed multi-party computation techniques is described.