Matches in SemOpenAlex for { <https://semopenalex.org/work/W4226176886> ?p ?o ?g. }
Showing items 1 to 86 of
86
with 100 items per page.
- W4226176886 abstract "Architecture reverse engineering has become an emerging attack against deep neural network (DNN) implementations. Several prior works have utilized side-channel leakage to recover the model architecture while the an DNN is executing on a hardware acceleration platform. In this work, we target an open-source deep-learning accelerator, Versatile Tensor Accelerator (VTA), and utilize electromagnetic (EM) side-channel leakage to comprehensively learn the association between DNN architecture configurations and EM emanations. We also consider the holistic system–including the low-level tensor program code of the VTA accelerator on a Xilinx FPGA, and explore the effect of such low-level configurations on the EM leakage. Our study demonstrates that both the optimization and configuration of tensor programs will affect the EM side-channel leakage.Gaining knowledge of the association between low-level tensor program and the EM emanations, we propose NNReArch, a lightweight tensor program scheduling framework against side-channel-based DNN model architecture reverse engineering. Specifically, NNReArch targets reshaping the EM traces of different DNN operators, through scheduling the tensor program execution of the DNN model so as to confuse the adversary. NNReArch is a comprehensive protection framework supporting two modes, a balanced mode that strikes a balance between the DNN model confidentiality and execution performance, and a secure mode where the most secure setting is chosen. We implement and evaluate the proposed framework on the open-source VTA with state-of-the-art DNN architectures. The experimental results demonstrate that NNReArch can efficiently enhance the model architecture security with a small performance overhead. In addition, the proposed obfuscation technique makes reverse engineering of the DNN architecture significantly harder." @default.
- W4226176886 created "2022-05-05" @default.
- W4226176886 creator A5020478086 @default.
- W4226176886 creator A5023765976 @default.
- W4226176886 creator A5041001467 @default.
- W4226176886 creator A5045682482 @default.
- W4226176886 creator A5083131515 @default.
- W4226176886 date "2022-05-15" @default.
- W4226176886 modified "2023-10-16" @default.
- W4226176886 title "NNReArch: A Tensor Program Scheduling Framework Against Neural Network Architecture Reverse Engineering" @default.
- W4226176886 cites W2442974303 @default.
- W4226176886 cites W2522548197 @default.
- W4226176886 cites W2606722458 @default.
- W4226176886 cites W3102836279 @default.
- W4226176886 cites W3114482311 @default.
- W4226176886 cites W3149866969 @default.
- W4226176886 cites W3166251174 @default.
- W4226176886 cites W3213911739 @default.
- W4226176886 cites W4200155238 @default.
- W4226176886 doi "https://doi.org/10.1109/fccm53951.2022.9786112" @default.
- W4226176886 hasPublicationYear "2022" @default.
- W4226176886 type Work @default.
- W4226176886 citedByCount "1" @default.
- W4226176886 countsByYear W42261768862022 @default.
- W4226176886 crossrefType "proceedings-article" @default.
- W4226176886 hasAuthorship W4226176886A5020478086 @default.
- W4226176886 hasAuthorship W4226176886A5023765976 @default.
- W4226176886 hasAuthorship W4226176886A5041001467 @default.
- W4226176886 hasAuthorship W4226176886A5045682482 @default.
- W4226176886 hasAuthorship W4226176886A5083131515 @default.
- W4226176886 hasBestOaLocation W42261768862 @default.
- W4226176886 hasConcept C108583219 @default.
- W4226176886 hasConcept C111919701 @default.
- W4226176886 hasConcept C113775141 @default.
- W4226176886 hasConcept C118524514 @default.
- W4226176886 hasConcept C123657996 @default.
- W4226176886 hasConcept C127413603 @default.
- W4226176886 hasConcept C142362112 @default.
- W4226176886 hasConcept C149635348 @default.
- W4226176886 hasConcept C153349607 @default.
- W4226176886 hasConcept C154945302 @default.
- W4226176886 hasConcept C178489894 @default.
- W4226176886 hasConcept C206729178 @default.
- W4226176886 hasConcept C207850805 @default.
- W4226176886 hasConcept C21547014 @default.
- W4226176886 hasConcept C38652104 @default.
- W4226176886 hasConcept C41008148 @default.
- W4226176886 hasConcept C42935608 @default.
- W4226176886 hasConcept C49289754 @default.
- W4226176886 hasConcept C50644808 @default.
- W4226176886 hasConceptScore W4226176886C108583219 @default.
- W4226176886 hasConceptScore W4226176886C111919701 @default.
- W4226176886 hasConceptScore W4226176886C113775141 @default.
- W4226176886 hasConceptScore W4226176886C118524514 @default.
- W4226176886 hasConceptScore W4226176886C123657996 @default.
- W4226176886 hasConceptScore W4226176886C127413603 @default.
- W4226176886 hasConceptScore W4226176886C142362112 @default.
- W4226176886 hasConceptScore W4226176886C149635348 @default.
- W4226176886 hasConceptScore W4226176886C153349607 @default.
- W4226176886 hasConceptScore W4226176886C154945302 @default.
- W4226176886 hasConceptScore W4226176886C178489894 @default.
- W4226176886 hasConceptScore W4226176886C206729178 @default.
- W4226176886 hasConceptScore W4226176886C207850805 @default.
- W4226176886 hasConceptScore W4226176886C21547014 @default.
- W4226176886 hasConceptScore W4226176886C38652104 @default.
- W4226176886 hasConceptScore W4226176886C41008148 @default.
- W4226176886 hasConceptScore W4226176886C42935608 @default.
- W4226176886 hasConceptScore W4226176886C49289754 @default.
- W4226176886 hasConceptScore W4226176886C50644808 @default.
- W4226176886 hasLocation W42261768861 @default.
- W4226176886 hasLocation W42261768862 @default.
- W4226176886 hasOpenAccess W4226176886 @default.
- W4226176886 hasPrimaryLocation W42261768861 @default.
- W4226176886 hasRelatedWork W2063534976 @default.
- W4226176886 hasRelatedWork W2116502997 @default.
- W4226176886 hasRelatedWork W2284838239 @default.
- W4226176886 hasRelatedWork W265063938 @default.
- W4226176886 hasRelatedWork W2794898833 @default.
- W4226176886 hasRelatedWork W2932459076 @default.
- W4226176886 hasRelatedWork W2995926156 @default.
- W4226176886 hasRelatedWork W3133757386 @default.
- W4226176886 hasRelatedWork W3195179447 @default.
- W4226176886 hasRelatedWork W4361251788 @default.
- W4226176886 isParatext "false" @default.
- W4226176886 isRetracted "false" @default.
- W4226176886 workType "article" @default.