Syft 0.5: A Platform for Universally Deployable Structured Transparency
Abstract
We present Syft 0.5, a general-purpose framework that combines a core group of privacy-enhancing technologies that facilitate a universal set of structured transparency systems. This framework is demonstrated through the design and implementation of a novel privacy-preserving inference information flow where we pass homomorphically encrypted activation signals through a split neural network for inference. We show that splitting the model further up the computation chain significantly reduces the computation time of inference and the payload size of activation signals at the cost of model secrecy. We evaluate our proposed flow with respect to its provision of the core structural transparency principles.
Cite
@article{arxiv.2104.12385,
title = {Syft 0.5: A Platform for Universally Deployable Structured Transparency},
author = {Adam James Hall and Madhava Jay and Tudor Cebere and Bogdan Cebere and Koen Lennart van der Veen and George Muraru and Tongye Xu and Patrick Cason and William Abramson and Ayoub Benaissa and Chinmay Shah and Alan Aboudib and Théo Ryffel and Kritika Prakash and Tom Titcombe and Varun Kumar Khare and Maddie Shang and Ionesio Junior and Animesh Gupta and Jason Paumier and Nahua Kang and Vova Manannikov and Andrew Trask},
journal= {arXiv preprint arXiv:2104.12385},
year = {2021}
}
Comments
ICLR 2021 Workshop on Distributed and Private Machine Learning (DPML 2021)