DC Field | Value | Language |
---|---|---|
dc.contributor.author | Lee, Hyunhoon | - |
dc.contributor.author | Lee, Youngjoo | - |
dc.date.accessioned | 2024-02-27T10:01:06Z | - |
dc.date.available | 2024-02-27T10:01:06Z | - |
dc.date.created | 2023-12-11 | - |
dc.date.issued | 2023-09 | - |
dc.identifier.uri | https://oasis.postech.ac.kr/handle/2014.oak/120429 | - |
dc.description.abstract | Homomorphic encryption (HE) based on the CKKS scheme is a promising candidate for implementing privacy-preserving deep neural networks (PP-DNN) by performing operations directly on the encrypted data. However, due to the computational complexity of HE operation, even simple PP-DNNs require a huge amount of processing time. In order to reduce the processing time of PP-DNN, in this paper, we present an innovative, low-latency model optimization solution for PP-DNNs. Our proposed low-latency model optimization solution exploits second-order polynomials that approximate original activation functions, ensuring low-latency and accurate DNN performance. To further reduce the processing latency of PP-DNNs, we introduce the coefficient absorbing technique and a masking convolution for convolutional layers. The experimental results show that the proposed solution constructs bootstrapping-free PP-DNN and reduces the inference latency of CKKS-based ResNet-34 by 35% in the CIFAR-100 dataset and ResNet-32 by 77% in the CIFAR-10 dataset compared to previous approaches while maintaining the same level of inference accuracy. Moreover, through the layer-wise latency analysis, we show the efficacy of our approaches, and through validation in various scenarios, we demonstrate the generality of our methods. | - |
dc.language | English | - |
dc.publisher | Institute of Electrical and Electronics Engineers Inc. | - |
dc.relation.isPartOf | IEEE Access | - |
dc.title | Optimizations of Privacy-Preserving DNN for Low-Latency Inference on Encrypted Data | - |
dc.type | Article | - |
dc.identifier.doi | 10.1109/ACCESS.2023.3318433 | - |
dc.type.rims | ART | - |
dc.identifier.bibliographicCitation | IEEE Access, v.11, pp.104775 - 104788 | - |
dc.identifier.wosid | 001080721700001 | - |
dc.citation.endPage | 104788 | - |
dc.citation.startPage | 104775 | - |
dc.citation.title | IEEE Access | - |
dc.citation.volume | 11 | - |
dc.contributor.affiliatedAuthor | Lee, Hyunhoon | - |
dc.contributor.affiliatedAuthor | Lee, Youngjoo | - |
dc.identifier.scopusid | 2-s2.0-85173041072 | - |
dc.description.journalClass | 1 | - |
dc.description.journalClass | 1 | - |
dc.description.isOpenAccess | Y | - |
dc.type.docType | Article | - |
dc.subject.keywordPlus | HOMOMORPHIC ENCRYPTION | - |
dc.subject.keywordAuthor | RNS-CKKS | - |
dc.subject.keywordAuthor | convolutional neural network | - |
dc.subject.keywordAuthor | homomorphic encryption | - |
dc.subject.keywordAuthor | privacy preserving neural network | - |
dc.relation.journalWebOfScienceCategory | Computer Science, Information Systems | - |
dc.relation.journalWebOfScienceCategory | Engineering, Electrical & Electronic | - |
dc.relation.journalWebOfScienceCategory | Telecommunications | - |
dc.description.journalRegisteredClass | scie | - |
dc.description.journalRegisteredClass | scopus | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
library@postech.ac.kr Tel: 054-279-2548
Copyrights © by 2017 Pohang University of Science ad Technology All right reserved.