ProteiNN: Privacy-preserving One-to-Many Neural Network Classifications

Beyza Bozdemir, Orhan Ermis, Melek Önen

2020

Abstract

In this work, we propose ProteiNN, a privacy-preserving neural network classification solution in a one-to-many scenario whereby one model provider outsources a machine learning model to the cloud server for its many different customers, and wishes to keep the model confidential while controlling its use. On the other hand, these customers take advantage of this machine learning model without revealing their sensitive inputs and the corresponding results. The solution employs homomorphic proxy re-encryption and a simple additive encryption to ensure the privacy of customers’ inputs and results against the model provider and the cloud server, and to give the control on the privacy and use of the model to the model provider. A detailed security analysis considering potential collusions among different players is provided.

Paper Citation

in Harvard Style

Bozdemir B., Ermis O. and Önen M. (2020). ProteiNN: Privacy-preserving One-to-Many Neural Network Classifications.In Proceedings of the 17th International Joint Conference on e-Business and Telecommunications - Volume 3: SECRYPT, ISBN 978-989-758-446-6, pages 397-404. DOI: 10.5220/0009829603970404

in Bibtex Style

@conference{secrypt20,
author={Beyza Bozdemir and Orhan Ermis and Melek Önen},
title={ProteiNN: Privacy-preserving One-to-Many Neural Network Classifications},
booktitle={Proceedings of the 17th International Joint Conference on e-Business and Telecommunications - Volume 3: SECRYPT,},
year={2020},
pages={397-404},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009829603970404},
isbn={978-989-758-446-6},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Joint Conference on e-Business and Telecommunications - Volume 3: SECRYPT,
TI - ProteiNN: Privacy-preserving One-to-Many Neural Network Classifications
SN - 978-989-758-446-6
AU - Bozdemir B.
AU - Ermis O.
AU - Önen M.
PY - 2020
SP - 397
EP - 404
DO - 10.5220/0009829603970404