Matches in SemOpenAlex for { <https://semopenalex.org/work/W4281619157> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4281619157 abstract "Deep Convolutional Neural Networks (DCNNs) have revolutionized and improved many aspects of modern life. However, these models are increasingly more complex, and training them to perform at desirable levels is difficult undertaking; hence, the trained parameters represent a valuable intellectual property (IP) asset which a motivated attacker may wish to steal. To better protect the IP, we propose a method of lightweight input obfuscation that is undone prior to inference, where input data is obfuscated in order to use the model to specification. Without using the correct key and unlocking sequence, the accuracy of the classifier is reduced to a random guess, thus protecting the input/output interface and mitigating model extraction attacks which rely on such access. We evaluate the system using a VGG-16 network trained on CIFAR-10, and demonstrate that with an incorrect deobfuscation key or sequence, the classification accuracy drops to a random guess, with an inference timing overhead of 4.4% on an Nvidia-based evaluation platform. The system avoids the costs associated with retraining and has no impact on model accuracy for authorized users." @default.
- W4281619157 created "2022-06-13" @default.
- W4281619157 creator A5049193989 @default.
- W4281619157 creator A5067475716 @default.
- W4281619157 date "2022-06-06" @default.
- W4281619157 modified "2023-09-27" @default.
- W4281619157 title "Protecting Deep Neural Network Intellectual Property with Architecture-Agnostic Input Obfuscation" @default.
- W4281619157 cites W2016380472 @default.
- W4281619157 cites W2051267297 @default.
- W4281619157 cites W2053637704 @default.
- W4281619157 cites W2311607323 @default.
- W4281619157 cites W2342408547 @default.
- W4281619157 cites W2561981131 @default.
- W4281619157 cites W2588204304 @default.
- W4281619157 cites W2623207919 @default.
- W4281619157 cites W2794284562 @default.
- W4281619157 cites W2808195004 @default.
- W4281619157 cites W2964335264 @default.
- W4281619157 cites W2971122390 @default.
- W4281619157 cites W2973414778 @default.
- W4281619157 cites W2995164854 @default.
- W4281619157 cites W3047375952 @default.
- W4281619157 cites W3092557510 @default.
- W4281619157 cites W3114482311 @default.
- W4281619157 cites W3118138297 @default.
- W4281619157 cites W3134815184 @default.
- W4281619157 cites W3141043040 @default.
- W4281619157 cites W3158697642 @default.
- W4281619157 cites W3183268554 @default.
- W4281619157 cites W2974645258 @default.
- W4281619157 doi "https://doi.org/10.1145/3526241.3530386" @default.
- W4281619157 hasPublicationYear "2022" @default.
- W4281619157 type Work @default.
- W4281619157 citedByCount "0" @default.
- W4281619157 crossrefType "proceedings-article" @default.
- W4281619157 hasAuthorship W4281619157A5049193989 @default.
- W4281619157 hasAuthorship W4281619157A5067475716 @default.
- W4281619157 hasConcept C108583219 @default.
- W4281619157 hasConcept C119857082 @default.
- W4281619157 hasConcept C124101348 @default.
- W4281619157 hasConcept C154945302 @default.
- W4281619157 hasConcept C26517878 @default.
- W4281619157 hasConcept C2776214188 @default.
- W4281619157 hasConcept C38652104 @default.
- W4281619157 hasConcept C40305131 @default.
- W4281619157 hasConcept C41008148 @default.
- W4281619157 hasConcept C50644808 @default.
- W4281619157 hasConcept C81363708 @default.
- W4281619157 hasConcept C95623464 @default.
- W4281619157 hasConceptScore W4281619157C108583219 @default.
- W4281619157 hasConceptScore W4281619157C119857082 @default.
- W4281619157 hasConceptScore W4281619157C124101348 @default.
- W4281619157 hasConceptScore W4281619157C154945302 @default.
- W4281619157 hasConceptScore W4281619157C26517878 @default.
- W4281619157 hasConceptScore W4281619157C2776214188 @default.
- W4281619157 hasConceptScore W4281619157C38652104 @default.
- W4281619157 hasConceptScore W4281619157C40305131 @default.
- W4281619157 hasConceptScore W4281619157C41008148 @default.
- W4281619157 hasConceptScore W4281619157C50644808 @default.
- W4281619157 hasConceptScore W4281619157C81363708 @default.
- W4281619157 hasConceptScore W4281619157C95623464 @default.
- W4281619157 hasLocation W42816191571 @default.
- W4281619157 hasOpenAccess W4281619157 @default.
- W4281619157 hasPrimaryLocation W42816191571 @default.
- W4281619157 hasRelatedWork W2337926734 @default.
- W4281619157 hasRelatedWork W2963058055 @default.
- W4281619157 hasRelatedWork W2986507176 @default.
- W4281619157 hasRelatedWork W3158264953 @default.
- W4281619157 hasRelatedWork W4293087779 @default.
- W4281619157 hasRelatedWork W4310989423 @default.
- W4281619157 hasRelatedWork W4311257506 @default.
- W4281619157 hasRelatedWork W4320802194 @default.
- W4281619157 hasRelatedWork W4366224123 @default.
- W4281619157 hasRelatedWork W564581980 @default.
- W4281619157 isParatext "false" @default.
- W4281619157 isRetracted "false" @default.
- W4281619157 workType "article" @default.