SecureBP from Homomorphic Encryption

SecureBP from Homomorphic Encryption

We present a secure backpropagation neural network training model (SecureBP), which allows a neural network to be trained while retaining the confidentiality of the training data, based on the homomorphic encryption scheme. We make two contributions. The first one is to introduce a method to find a...

Full description

Bibliographic Details
Main Authors: Qinju Liu, Xianhui Lu, Fucai Luo, Shuai Zhou, Jingnan He, Kunpeng Wang
Format: Article
Language:English
Published: Hindawi-Wiley 2020-01-01
Series:Security and Communication Networks
Online Access:http://dx.doi.org/10.1155/2020/5328059
id doaj-4775bf9024494f60a4800819d20be875
record_format Article
spelling doaj-4775bf9024494f60a4800819d20be8752020-11-25T02:31:32ZengHindawi-WileySecurity and Communication Networks1939-01141939-01222020-01-01202010.1155/2020/53280595328059SecureBP from Homomorphic EncryptionQinju Liu0Xianhui Lu1Fucai Luo2Shuai Zhou3Jingnan He4Kunpeng Wang5State Key Laboratory of Information Security, Institute of Information Engineering, CAS, Beijing 100093, ChinaState Key Laboratory of Information Security, Institute of Information Engineering, CAS, Beijing 100093, ChinaState Key Laboratory of Information Security, Institute of Information Engineering, CAS, Beijing 100093, ChinaFaculty of Engineering and Information Technology Engineering, University of Technology Sydney, Ultimo, NSW 2007, AustraliaState Key Laboratory of Information Security, Institute of Information Engineering, CAS, Beijing 100093, ChinaState Key Laboratory of Information Security, Institute of Information Engineering, CAS, Beijing 100093, ChinaWe present a secure backpropagation neural network training model (SecureBP), which allows a neural network to be trained while retaining the confidentiality of the training data, based on the homomorphic encryption scheme. We make two contributions. The first one is to introduce a method to find a more accurate and numerically stable polynomial approximation of functions in a certain interval. The second one is to find a strategy of refreshing ciphertext during training, which keeps the order of magnitude of noise at O˜e33.http://dx.doi.org/10.1155/2020/5328059
collection DOAJ
language English
format Article
sources DOAJ
author Qinju Liu
Xianhui Lu
Fucai Luo
Shuai Zhou
Jingnan He
Kunpeng Wang
spellingShingle Qinju Liu
Xianhui Lu
Fucai Luo
Shuai Zhou
Jingnan He
Kunpeng Wang
SecureBP from Homomorphic Encryption
Security and Communication Networks
author_facet Qinju Liu
Xianhui Lu
Fucai Luo
Shuai Zhou
Jingnan He
Kunpeng Wang
author_sort Qinju Liu
title SecureBP from Homomorphic Encryption
title_short SecureBP from Homomorphic Encryption
title_full SecureBP from Homomorphic Encryption
title_fullStr SecureBP from Homomorphic Encryption
title_full_unstemmed SecureBP from Homomorphic Encryption
title_sort securebp from homomorphic encryption
publisher Hindawi-Wiley
series Security and Communication Networks
issn 1939-0114
1939-0122
publishDate 2020-01-01
description We present a secure backpropagation neural network training model (SecureBP), which allows a neural network to be trained while retaining the confidentiality of the training data, based on the homomorphic encryption scheme. We make two contributions. The first one is to introduce a method to find a more accurate and numerically stable polynomial approximation of functions in a certain interval. The second one is to find a strategy of refreshing ciphertext during training, which keeps the order of magnitude of noise at O˜e33.
url http://dx.doi.org/10.1155/2020/5328059
work_keys_str_mv AT qinjuliu securebpfromhomomorphicencryption
AT xianhuilu securebpfromhomomorphicencryption
AT fucailuo securebpfromhomomorphicencryption
AT shuaizhou securebpfromhomomorphicencryption
AT jingnanhe securebpfromhomomorphicencryption
AT kunpengwang securebpfromhomomorphicencryption
_version_ 1715462317378895872

Similar Items