Secure Evaluation of Quantized Neural Networks

We investigate two questions in this paper: First, we ask to what extent “MPC friendly” models are already supported by major Machine Learning frameworks such as TensorFlow or PyTorch. Prior works provide protocols that only work on fixed-point integers and specialized activation functions, two aspe...

Full description

Bibliographic Details
Main Authors: Dalskov Anders, Escudero Daniel, Keller Marcel
Format: Article
Language:English
Published: Sciendo 2020-10-01
Series:Proceedings on Privacy Enhancing Technologies
Online Access:https://doi.org/10.2478/popets-2020-0077