Matches in SemOpenAlex for { <https://semopenalex.org/work/W3114482311> ?p ?o ?g. }
- W3114482311 abstract "Neural Network (NN) accelerators are currently widely deployed in various security-crucial scenarios, including image recognition, natural language processing and autonomous vehicles. Due to economic and privacy concerns, the hardware implementations of structures and designs inside NN accelerators are usually inaccessible to the public. However, these accelerators still tend to leak crucial information through Electromagnetic (EM) side channels in addition to timing and power information. In this paper, we propose an effective and efficient model stealing attack against current popular large-scale NN accelerators deployed on hardware platforms through side-channel information. Specifically, the proposed attack approach contains two stages: 1) Inferring the underlying network architecture through EM sidechannel information; 2) Estimating the parameters, especially the weights, through a margin-based, adversarial active learning method. The experimental results show that the proposed attack approach can accurately recover the large-scale NN through EM side-channel information leakages. Overall, our attack highlights the importance of masking EM traces for large-scale NN accelerators in real-world applications." @default.
- W3114482311 created "2021-01-05" @default.
- W3114482311 creator A5011038036 @default.
- W3114482311 creator A5017464942 @default.
- W3114482311 creator A5060297615 @default.
- W3114482311 creator A5087134798 @default.
- W3114482311 creator A5090083857 @default.
- W3114482311 date "2020-12-07" @default.
- W3114482311 modified "2023-10-16" @default.
- W3114482311 title "DeepEM: Deep Neural Networks Model Recovery through EM Side-Channel Information Leakage" @default.
- W3114482311 cites W1511843316 @default.
- W3114482311 cites W1557204831 @default.
- W3114482311 cites W1724890242 @default.
- W3114482311 cites W1885172175 @default.
- W3114482311 cites W1920702274 @default.
- W3114482311 cites W1998808035 @default.
- W3114482311 cites W2112796928 @default.
- W3114482311 cites W2171671120 @default.
- W3114482311 cites W2194775991 @default.
- W3114482311 cites W2585560244 @default.
- W3114482311 cites W2591924527 @default.
- W3114482311 cites W2603766943 @default.
- W3114482311 cites W2618530766 @default.
- W3114482311 cites W2724189024 @default.
- W3114482311 cites W2895160027 @default.
- W3114482311 cites W2949071206 @default.
- W3114482311 cites W2949388520 @default.
- W3114482311 cites W2953941229 @default.
- W3114482311 cites W2963303354 @default.
- W3114482311 cites W2964149421 @default.
- W3114482311 cites W3007318395 @default.
- W3114482311 cites W3102169921 @default.
- W3114482311 cites W3149866969 @default.
- W3114482311 cites W2974645258 @default.
- W3114482311 doi "https://doi.org/10.1109/host45689.2020.9300274" @default.
- W3114482311 hasPublicationYear "2020" @default.
- W3114482311 type Work @default.
- W3114482311 sameAs 3114482311 @default.
- W3114482311 citedByCount "41" @default.
- W3114482311 countsByYear W31144823112020 @default.
- W3114482311 countsByYear W31144823112021 @default.
- W3114482311 countsByYear W31144823112022 @default.
- W3114482311 countsByYear W31144823112023 @default.
- W3114482311 crossrefType "proceedings-article" @default.
- W3114482311 hasAuthorship W3114482311A5011038036 @default.
- W3114482311 hasAuthorship W3114482311A5017464942 @default.
- W3114482311 hasAuthorship W3114482311A5060297615 @default.
- W3114482311 hasAuthorship W3114482311A5087134798 @default.
- W3114482311 hasAuthorship W3114482311A5090083857 @default.
- W3114482311 hasConcept C108583219 @default.
- W3114482311 hasConcept C113775141 @default.
- W3114482311 hasConcept C115903868 @default.
- W3114482311 hasConcept C119857082 @default.
- W3114482311 hasConcept C127162648 @default.
- W3114482311 hasConcept C139719470 @default.
- W3114482311 hasConcept C142362112 @default.
- W3114482311 hasConcept C153349607 @default.
- W3114482311 hasConcept C154945302 @default.
- W3114482311 hasConcept C162324750 @default.
- W3114482311 hasConcept C178489894 @default.
- W3114482311 hasConcept C26713055 @default.
- W3114482311 hasConcept C2777042071 @default.
- W3114482311 hasConcept C2777402240 @default.
- W3114482311 hasConcept C2779201187 @default.
- W3114482311 hasConcept C31258907 @default.
- W3114482311 hasConcept C38652104 @default.
- W3114482311 hasConcept C41008148 @default.
- W3114482311 hasConcept C49289754 @default.
- W3114482311 hasConcept C50644808 @default.
- W3114482311 hasConcept C774472 @default.
- W3114482311 hasConceptScore W3114482311C108583219 @default.
- W3114482311 hasConceptScore W3114482311C113775141 @default.
- W3114482311 hasConceptScore W3114482311C115903868 @default.
- W3114482311 hasConceptScore W3114482311C119857082 @default.
- W3114482311 hasConceptScore W3114482311C127162648 @default.
- W3114482311 hasConceptScore W3114482311C139719470 @default.
- W3114482311 hasConceptScore W3114482311C142362112 @default.
- W3114482311 hasConceptScore W3114482311C153349607 @default.
- W3114482311 hasConceptScore W3114482311C154945302 @default.
- W3114482311 hasConceptScore W3114482311C162324750 @default.
- W3114482311 hasConceptScore W3114482311C178489894 @default.
- W3114482311 hasConceptScore W3114482311C26713055 @default.
- W3114482311 hasConceptScore W3114482311C2777042071 @default.
- W3114482311 hasConceptScore W3114482311C2777402240 @default.
- W3114482311 hasConceptScore W3114482311C2779201187 @default.
- W3114482311 hasConceptScore W3114482311C31258907 @default.
- W3114482311 hasConceptScore W3114482311C38652104 @default.
- W3114482311 hasConceptScore W3114482311C41008148 @default.
- W3114482311 hasConceptScore W3114482311C49289754 @default.
- W3114482311 hasConceptScore W3114482311C50644808 @default.
- W3114482311 hasConceptScore W3114482311C774472 @default.
- W3114482311 hasFunder F4320337345 @default.
- W3114482311 hasLocation W31144823111 @default.
- W3114482311 hasOpenAccess W3114482311 @default.
- W3114482311 hasPrimaryLocation W31144823111 @default.
- W3114482311 hasRelatedWork W1873852107 @default.
- W3114482311 hasRelatedWork W2399451741 @default.
- W3114482311 hasRelatedWork W2539268621 @default.
- W3114482311 hasRelatedWork W2586046457 @default.
- W3114482311 hasRelatedWork W2586683750 @default.