Matches in SemOpenAlex for { <https://semopenalex.org/work/W4366557487> ?p ?o ?g. }
Showing items 1 to 63 of
63
with 100 items per page.
- W4366557487 abstract "Homomorphic encryption (HE) enables calculating on encrypted data, which makes it possible to perform privacypreserving neural network inference. One disadvantage of this technique is that it is several orders of magnitudes slower than calculation on unencrypted data. Neural networks are commonly trained using floating-point, while most homomorphic encryption libraries calculate on integers, thus requiring a quantisation of the neural network. A straightforward approach would be to quantise to large integer sizes (e.g. 32 bit) to avoid large quantisation errors. In this work, we reduce the integer sizes of the networks, using quantisation-aware training, to allow more efficient computations. For the targeted MNIST architecture proposed by Badawi et al., we reduce the integer sizes by 33% without significant loss of accuracy, while for the CIFAR architecture, we can reduce the integer sizes by 43%. Implementing the resulting networks under the BFV homomorphic encryption scheme using SEAL, we could reduce the execution time of an MNIST neural network by 80% and by 40% for a CIFAR neural network." @default.
- W4366557487 created "2023-04-22" @default.
- W4366557487 creator A5017161705 @default.
- W4366557487 creator A5018532230 @default.
- W4366557487 creator A5019332272 @default.
- W4366557487 creator A5048759552 @default.
- W4366557487 creator A5082347771 @default.
- W4366557487 date "2023-04-19" @default.
- W4366557487 modified "2023-09-27" @default.
- W4366557487 title "Neural Network Quantisation for Faster Homomorphic Encryption" @default.
- W4366557487 doi "https://doi.org/10.1109/iolts59296.2023.10224890" @default.
- W4366557487 hasPublicationYear "2023" @default.
- W4366557487 type Work @default.
- W4366557487 citedByCount "0" @default.
- W4366557487 crossrefType "posted-content" @default.
- W4366557487 hasAuthorship W4366557487A5017161705 @default.
- W4366557487 hasAuthorship W4366557487A5018532230 @default.
- W4366557487 hasAuthorship W4366557487A5019332272 @default.
- W4366557487 hasAuthorship W4366557487A5048759552 @default.
- W4366557487 hasAuthorship W4366557487A5082347771 @default.
- W4366557487 hasBestOaLocation W43665574871 @default.
- W4366557487 hasConcept C11413529 @default.
- W4366557487 hasConcept C148730421 @default.
- W4366557487 hasConcept C154945302 @default.
- W4366557487 hasConcept C158338273 @default.
- W4366557487 hasConcept C190502265 @default.
- W4366557487 hasConcept C199360897 @default.
- W4366557487 hasConcept C2776214188 @default.
- W4366557487 hasConcept C31258907 @default.
- W4366557487 hasConcept C41008148 @default.
- W4366557487 hasConcept C50644808 @default.
- W4366557487 hasConcept C80444323 @default.
- W4366557487 hasConcept C84211073 @default.
- W4366557487 hasConcept C97137487 @default.
- W4366557487 hasConceptScore W4366557487C11413529 @default.
- W4366557487 hasConceptScore W4366557487C148730421 @default.
- W4366557487 hasConceptScore W4366557487C154945302 @default.
- W4366557487 hasConceptScore W4366557487C158338273 @default.
- W4366557487 hasConceptScore W4366557487C190502265 @default.
- W4366557487 hasConceptScore W4366557487C199360897 @default.
- W4366557487 hasConceptScore W4366557487C2776214188 @default.
- W4366557487 hasConceptScore W4366557487C31258907 @default.
- W4366557487 hasConceptScore W4366557487C41008148 @default.
- W4366557487 hasConceptScore W4366557487C50644808 @default.
- W4366557487 hasConceptScore W4366557487C80444323 @default.
- W4366557487 hasConceptScore W4366557487C84211073 @default.
- W4366557487 hasConceptScore W4366557487C97137487 @default.
- W4366557487 hasLocation W43665574871 @default.
- W4366557487 hasOpenAccess W4366557487 @default.
- W4366557487 hasPrimaryLocation W43665574871 @default.
- W4366557487 hasRelatedWork W1623766202 @default.
- W4366557487 hasRelatedWork W2340393497 @default.
- W4366557487 hasRelatedWork W2395982605 @default.
- W4366557487 hasRelatedWork W2949060133 @default.
- W4366557487 hasRelatedWork W2966207885 @default.
- W4366557487 hasRelatedWork W3205328660 @default.
- W4366557487 hasRelatedWork W4229059147 @default.
- W4366557487 hasRelatedWork W4308654840 @default.
- W4366557487 hasRelatedWork W4362712035 @default.
- W4366557487 hasRelatedWork W4366557487 @default.
- W4366557487 isParatext "false" @default.
- W4366557487 isRetracted "false" @default.
- W4366557487 workType "article" @default.