Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287331609> ?p ?o ?g. }
Showing items 1 to 72 of
72
with 100 items per page.
- W4287331609 abstract "Homomorphic Encryption (HE), allowing computations on encrypted data (ciphertext) without decrypting it first, enables secure but prohibitively slow Convolutional Neural Network (CNN) inference for privacy-preserving applications in clouds. To reduce the inference latency, one approach is to pack multiple messages into a single ciphertext in order to reduce the number of ciphertexts and support massive parallelism of Homomorphic Multiply-Accumulate (HMA) operations between ciphertexts. Despite the faster HECNN inference, the mainstream packing schemes Dense Packing (DensePack) and Convolution Packing (ConvPack) introduce expensive rotation overhead, which prolongs the inference latency of HECNN for deeper and wider CNN architectures. In this paper, we propose a low-rank factorization method named FFConv dedicated to efficient ciphertext packing for reducing both the rotation overhead and HMA operations. FFConv approximates a d x d convolution layer with low-rank factorized convolutions, in which a d x d low-rank convolution with fewer channels is followed by a 1 x 1 convolution to restore the channels. The d x d low-rank convolution with DensePack leads to significantly reduced rotation operations, while the rotation overhead of 1 x 1 convolution with ConvPack is close to zero. To our knowledge, FFConv is the first work that is capable of reducing the rotation overhead incurred by DensePack and ConvPack simultaneously, without introducing additional special blocks into the HECNN inference pipeline. Compared to prior art LoLa and Falcon, our method reduces the inference latency by up to 88% and 21%, respectively, with comparable accuracy on MNIST and CIFAR-10." @default.
- W4287331609 created "2022-07-25" @default.
- W4287331609 creator A5015102287 @default.
- W4287331609 creator A5033224111 @default.
- W4287331609 creator A5036457690 @default.
- W4287331609 creator A5056053058 @default.
- W4287331609 creator A5058648309 @default.
- W4287331609 creator A5063447635 @default.
- W4287331609 creator A5087333350 @default.
- W4287331609 date "2021-02-05" @default.
- W4287331609 modified "2023-09-28" @default.
- W4287331609 title "FFConv: Fast Factorized Convolutional Neural Network Inference on Encrypted Data" @default.
- W4287331609 doi "https://doi.org/10.48550/arxiv.2102.03494" @default.
- W4287331609 hasPublicationYear "2021" @default.
- W4287331609 type Work @default.
- W4287331609 citedByCount "0" @default.
- W4287331609 crossrefType "posted-content" @default.
- W4287331609 hasAuthorship W4287331609A5015102287 @default.
- W4287331609 hasAuthorship W4287331609A5033224111 @default.
- W4287331609 hasAuthorship W4287331609A5036457690 @default.
- W4287331609 hasAuthorship W4287331609A5056053058 @default.
- W4287331609 hasAuthorship W4287331609A5058648309 @default.
- W4287331609 hasAuthorship W4287331609A5063447635 @default.
- W4287331609 hasAuthorship W4287331609A5087333350 @default.
- W4287331609 hasBestOaLocation W42873316091 @default.
- W4287331609 hasConcept C111919701 @default.
- W4287331609 hasConcept C11413529 @default.
- W4287331609 hasConcept C148730421 @default.
- W4287331609 hasConcept C154945302 @default.
- W4287331609 hasConcept C173608175 @default.
- W4287331609 hasConcept C2776214188 @default.
- W4287331609 hasConcept C2779960059 @default.
- W4287331609 hasConcept C31258907 @default.
- W4287331609 hasConcept C41008148 @default.
- W4287331609 hasConcept C45347329 @default.
- W4287331609 hasConcept C50644808 @default.
- W4287331609 hasConcept C74050887 @default.
- W4287331609 hasConcept C80444323 @default.
- W4287331609 hasConcept C81363708 @default.
- W4287331609 hasConcept C93974786 @default.
- W4287331609 hasConceptScore W4287331609C111919701 @default.
- W4287331609 hasConceptScore W4287331609C11413529 @default.
- W4287331609 hasConceptScore W4287331609C148730421 @default.
- W4287331609 hasConceptScore W4287331609C154945302 @default.
- W4287331609 hasConceptScore W4287331609C173608175 @default.
- W4287331609 hasConceptScore W4287331609C2776214188 @default.
- W4287331609 hasConceptScore W4287331609C2779960059 @default.
- W4287331609 hasConceptScore W4287331609C31258907 @default.
- W4287331609 hasConceptScore W4287331609C41008148 @default.
- W4287331609 hasConceptScore W4287331609C45347329 @default.
- W4287331609 hasConceptScore W4287331609C50644808 @default.
- W4287331609 hasConceptScore W4287331609C74050887 @default.
- W4287331609 hasConceptScore W4287331609C80444323 @default.
- W4287331609 hasConceptScore W4287331609C81363708 @default.
- W4287331609 hasConceptScore W4287331609C93974786 @default.
- W4287331609 hasLocation W42873316091 @default.
- W4287331609 hasLocation W42873316092 @default.
- W4287331609 hasOpenAccess W4287331609 @default.
- W4287331609 hasPrimaryLocation W42873316091 @default.
- W4287331609 hasRelatedWork W2014146283 @default.
- W4287331609 hasRelatedWork W2154234914 @default.
- W4287331609 hasRelatedWork W2177671734 @default.
- W4287331609 hasRelatedWork W2384710392 @default.
- W4287331609 hasRelatedWork W2502013096 @default.
- W4287331609 hasRelatedWork W2737488712 @default.
- W4287331609 hasRelatedWork W2913609182 @default.
- W4287331609 hasRelatedWork W291693969 @default.
- W4287331609 hasRelatedWork W2997006882 @default.
- W4287331609 hasRelatedWork W3031175382 @default.
- W4287331609 isParatext "false" @default.
- W4287331609 isRetracted "false" @default.
- W4287331609 workType "article" @default.