High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks
Due to legal restrictions or restrictions related to companies' internal information policies, businesses often do not trust sensitive information to public cloud providers. One of the mechanisms to ensure the security of sensitive data in clouds is homomorphic encryption. Privacy-preserving ne...
محفوظ في:
| المؤلفون الرئيسيون: | , , , , , |
|---|---|
| التنسيق: | Статья |
| اللغة: | English |
| منشور في: |
Pleiades Publishing
2024
|
| الموضوعات: | |
| الوصول للمادة أونلاين: | https://dspace.ncfu.ru/handle/123456789/29339 |
| الوسوم: |
إضافة وسم
لا توجد وسوم, كن أول من يضع وسما على هذه التسجيلة!
|
| id |
ir-123456789-29339 |
|---|---|
| record_format |
dspace |
| spelling |
ir-123456789-293392024-12-09T13:02:05Z High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks Lapina, M. A. Лапина, М. А. Shiriaev, E. M. Ширяев, Е. М. Babenko, M. G. Бабенко, М. Г. Convolutional neural networks Spatial complexity Cryptography Differential privacy High Speed Ho-momorphic encryptions Homomorphic-encryptions Information policy Legal restriction Multiplication algorithms Scalar multiplication Privacy preserving Neural-networks Convolution Due to legal restrictions or restrictions related to companies' internal information policies, businesses often do not trust sensitive information to public cloud providers. One of the mechanisms to ensure the security of sensitive data in clouds is homomorphic encryption. Privacy-preserving neural networks are used to design solutions that utilize neural networks under these conditions. They exploit the homomorphic encryption mechanism, thus enabling the security of commercial information in the cloud. The main deterrent to the use of privacy-preserving neural networks is the large computational and spatial complexity of the scalar multiplication algorithm, which is the basic algorithm for computing mathematical convolution. In this paper, we propose a scalar multiplication algorithm that reduces the spatial complexity from quadratic to linear, and reduces the computation time of scalar multiplication by a factor of 1.38. 2024-12-09T13:01:09Z 2024-12-09T13:01:09Z 2024 Статья Lapina, M.A., Shiriaev, E.M., Babenko, M.G., Istamov, I. High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks // Programming and Computer Software. - 2024. - 50 (6). - pp. 417-424. - DOI: 10.1134/S0361768824700282 https://dspace.ncfu.ru/handle/123456789/29339 en Programming and Computer Software application/pdf application/pdf Pleiades Publishing |
| institution |
СКФУ |
| collection |
Репозиторий |
| language |
English |
| topic |
Convolutional neural networks Spatial complexity Cryptography Differential privacy High Speed Ho-momorphic encryptions Homomorphic-encryptions Information policy Legal restriction Multiplication algorithms Scalar multiplication Privacy preserving Neural-networks Convolution |
| spellingShingle |
Convolutional neural networks Spatial complexity Cryptography Differential privacy High Speed Ho-momorphic encryptions Homomorphic-encryptions Information policy Legal restriction Multiplication algorithms Scalar multiplication Privacy preserving Neural-networks Convolution Lapina, M. A. Лапина, М. А. Shiriaev, E. M. Ширяев, Е. М. Babenko, M. G. Бабенко, М. Г. High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks |
| description |
Due to legal restrictions or restrictions related to companies' internal information policies, businesses often do not trust sensitive information to public cloud providers. One of the mechanisms to ensure the security of sensitive data in clouds is homomorphic encryption. Privacy-preserving neural networks are used to design solutions that utilize neural networks under these conditions. They exploit the homomorphic encryption mechanism, thus enabling the security of commercial information in the cloud. The main deterrent to the use of privacy-preserving neural networks is the large computational and spatial complexity of the scalar multiplication algorithm, which is the basic algorithm for computing mathematical convolution. In this paper, we propose a scalar multiplication algorithm that reduces the spatial complexity from quadratic to linear, and reduces the computation time of scalar multiplication by a factor of 1.38. |
| format |
Статья |
| author |
Lapina, M. A. Лапина, М. А. Shiriaev, E. M. Ширяев, Е. М. Babenko, M. G. Бабенко, М. Г. |
| author_facet |
Lapina, M. A. Лапина, М. А. Shiriaev, E. M. Ширяев, Е. М. Babenko, M. G. Бабенко, М. Г. |
| author_sort |
Lapina, M. A. |
| title |
High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks |
| title_short |
High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks |
| title_full |
High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks |
| title_fullStr |
High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks |
| title_full_unstemmed |
High-Speed Convolution Core Architecture for Privacy-Preserving Neural Networks |
| title_sort |
high-speed convolution core architecture for privacy-preserving neural networks |
| publisher |
Pleiades Publishing |
| publishDate |
2024 |
| url |
https://dspace.ncfu.ru/handle/123456789/29339 |
| work_keys_str_mv |
AT lapinama highspeedconvolutioncorearchitectureforprivacypreservingneuralnetworks AT lapinama highspeedconvolutioncorearchitectureforprivacypreservingneuralnetworks AT shiriaevem highspeedconvolutioncorearchitectureforprivacypreservingneuralnetworks AT širâevem highspeedconvolutioncorearchitectureforprivacypreservingneuralnetworks AT babenkomg highspeedconvolutioncorearchitectureforprivacypreservingneuralnetworks AT babenkomg highspeedconvolutioncorearchitectureforprivacypreservingneuralnetworks |
| _version_ |
1842245714811289600 |