Matches in SemOpenAlex for { <https://semopenalex.org/work/W3097924569> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W3097924569 abstract "Outsourcing machine learning inference creates a confidentiality dilemma: either the client has to trust the server with potentially sensitive input data, or the server has to share his commercially valuable model. Known remedies include homomorphic encryption, multi-party computation, or placing the entire model in a trusted enclave. None of these are suitable for large models. For two relevant use cases, we show that it is possible to keep all confidential model parameters in the last (dense) layers of deep neural networks. This allows us to split the model such that the confidential parts fit into a trusted enclave on the client side. We present the eNNclave toolchain to cut TensorFlow models at any layer, splitting them into public and enclaved layers. This preserves TensorFlow's performance optimizations and hardware support for public layers, while keeping the parameters of the enclaved layers private. Evaluations on several machine learning tasks spanning multiple domains show that fast inference is possible while keeping the sensitive model parameters confidential. Accuracy results are close to the baseline where all layers carry sensitive information and confirm our approach is practical." @default.
- W3097924569 created "2020-11-09" @default.
- W3097924569 creator A5029455813 @default.
- W3097924569 creator A5075251080 @default.
- W3097924569 date "2020-11-09" @default.
- W3097924569 modified "2023-09-26" @default.
- W3097924569 title "eNNclave" @default.
- W3097924569 cites W1851422430 @default.
- W3097924569 cites W2051267297 @default.
- W3097924569 cites W2117539524 @default.
- W3097924569 cites W2152161678 @default.
- W3097924569 cites W2165698076 @default.
- W3097924569 cites W2194775991 @default.
- W3097924569 cites W2295464403 @default.
- W3097924569 cites W2325939864 @default.
- W3097924569 cites W2535690855 @default.
- W3097924569 cites W2565439473 @default.
- W3097924569 cites W2603766943 @default.
- W3097924569 cites W2701059868 @default.
- W3097924569 cites W2794357950 @default.
- W3097924569 cites W2947718647 @default.
- W3097924569 cites W2963311060 @default.
- W3097924569 cites W2963857521 @default.
- W3097924569 cites W2973232880 @default.
- W3097924569 cites W3015806656 @default.
- W3097924569 doi "https://doi.org/10.1145/3411508.3421376" @default.
- W3097924569 hasPublicationYear "2020" @default.
- W3097924569 type Work @default.
- W3097924569 sameAs 3097924569 @default.
- W3097924569 citedByCount "6" @default.
- W3097924569 countsByYear W30979245692021 @default.
- W3097924569 countsByYear W30979245692022 @default.
- W3097924569 countsByYear W30979245692023 @default.
- W3097924569 crossrefType "proceedings-article" @default.
- W3097924569 hasAuthorship W3097924569A5029455813 @default.
- W3097924569 hasAuthorship W3097924569A5075251080 @default.
- W3097924569 hasConcept C111472728 @default.
- W3097924569 hasConcept C111919701 @default.
- W3097924569 hasConcept C11413529 @default.
- W3097924569 hasConcept C119857082 @default.
- W3097924569 hasConcept C138885662 @default.
- W3097924569 hasConcept C148730421 @default.
- W3097924569 hasConcept C154945302 @default.
- W3097924569 hasConcept C158338273 @default.
- W3097924569 hasConcept C17744445 @default.
- W3097924569 hasConcept C178790620 @default.
- W3097924569 hasConcept C185592680 @default.
- W3097924569 hasConcept C199539241 @default.
- W3097924569 hasConcept C2776214188 @default.
- W3097924569 hasConcept C2777062904 @default.
- W3097924569 hasConcept C2777904410 @default.
- W3097924569 hasConcept C2778496695 @default.
- W3097924569 hasConcept C2779227376 @default.
- W3097924569 hasConcept C38652104 @default.
- W3097924569 hasConcept C41008148 @default.
- W3097924569 hasConcept C45374587 @default.
- W3097924569 hasConcept C46934059 @default.
- W3097924569 hasConcept C50644808 @default.
- W3097924569 hasConcept C71745522 @default.
- W3097924569 hasConceptScore W3097924569C111472728 @default.
- W3097924569 hasConceptScore W3097924569C111919701 @default.
- W3097924569 hasConceptScore W3097924569C11413529 @default.
- W3097924569 hasConceptScore W3097924569C119857082 @default.
- W3097924569 hasConceptScore W3097924569C138885662 @default.
- W3097924569 hasConceptScore W3097924569C148730421 @default.
- W3097924569 hasConceptScore W3097924569C154945302 @default.
- W3097924569 hasConceptScore W3097924569C158338273 @default.
- W3097924569 hasConceptScore W3097924569C17744445 @default.
- W3097924569 hasConceptScore W3097924569C178790620 @default.
- W3097924569 hasConceptScore W3097924569C185592680 @default.
- W3097924569 hasConceptScore W3097924569C199539241 @default.
- W3097924569 hasConceptScore W3097924569C2776214188 @default.
- W3097924569 hasConceptScore W3097924569C2777062904 @default.
- W3097924569 hasConceptScore W3097924569C2777904410 @default.
- W3097924569 hasConceptScore W3097924569C2778496695 @default.
- W3097924569 hasConceptScore W3097924569C2779227376 @default.
- W3097924569 hasConceptScore W3097924569C38652104 @default.
- W3097924569 hasConceptScore W3097924569C41008148 @default.
- W3097924569 hasConceptScore W3097924569C45374587 @default.
- W3097924569 hasConceptScore W3097924569C46934059 @default.
- W3097924569 hasConceptScore W3097924569C50644808 @default.
- W3097924569 hasConceptScore W3097924569C71745522 @default.
- W3097924569 hasLocation W30979245691 @default.
- W3097924569 hasOpenAccess W3097924569 @default.
- W3097924569 hasPrimaryLocation W30979245691 @default.
- W3097924569 hasRelatedWork W2097184218 @default.
- W3097924569 hasRelatedWork W2302221905 @default.
- W3097924569 hasRelatedWork W2383950765 @default.
- W3097924569 hasRelatedWork W2744788180 @default.
- W3097924569 hasRelatedWork W2759141714 @default.
- W3097924569 hasRelatedWork W2807659734 @default.
- W3097924569 hasRelatedWork W3181110568 @default.
- W3097924569 hasRelatedWork W3184904781 @default.
- W3097924569 hasRelatedWork W4308654840 @default.
- W3097924569 hasRelatedWork W1629725936 @default.
- W3097924569 isParatext "false" @default.
- W3097924569 isRetracted "false" @default.
- W3097924569 magId "3097924569" @default.
- W3097924569 workType "article" @default.