Efficient Evaluation of Activation Functions over Encrypted Data

Download PDF
Open Webpage

Patricia Thaine, Sergey Gorbunov, Gerald Penn

Published: 2019, Last Modified: 27 Sept 2024IEEE Symposium on Security and Privacy Workshops 2019EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: We describe a method for approximating any bounded activation function given encrypted input data. The utility of our method is exemplified by simulating it within two typical machine learning tasks: namely, a Variational Autoencoder that learns a latent representation of MNIST data, and an MNIST image classifier.