FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3113404214> ?p ?o ?g. }

• W3113404214 endingPage "1942" @default.
• W3113404214 startingPage "1928" @default.
• W3113404214 abstract "Deep neural network (DNN) accelerators overcome the power and memory walls for executing neural-net models locally on edge-computing devices to support sophisticated AI applications. The advocacy of “model once, run optimized anywhere” paradigm introduces potential new security threat to edge intelligence that is methodologically different from the well-known adversarial examples. Existing adversarial examples modify the input samples presented to an AI application either digitally or physically to cause a misclassification. Nevertheless, these input-based perturbations are not robust or surreptitious on multi-view target. To generate a good adversarial example for misclassifying a real-world target of variational viewing angle, lighting and distance, a decent number of target’s samples are required to extract the rare anomalies that can cross the decision boundary. The feasible perturbations are substantial and visually perceptible. In this paper, we propose a new glitch injection attack on DNN accelerator that is capable of misclassifying a target under variational viewpoints. The glitches injected into the computation clock signal induce transitory but disruptive errors in the intermediate results of the multiply-and-accumulate (MAC) operations. The attack pattern for each target of interest consists of sparse instantaneous glitches, which can be derived from just one sample of the target. Two modes of attack patterns are derived, and their effectiveness are demonstrated on four representative ImageNet models implemented on the Deep-learning Processing Unit (DPU) of FPGA edge and its DNN development toolchain. The attack success rates are evaluated on 118 objects in 61 diverse sensing conditions, including 25 viewing angles (−60° to 60°), 24 illumination directions and 12 color temperatures. In the covert mode, the success rates of our attack exceed existing stealthy adversarial examples by more than 16.3%, with only two glitches injected into ten thousands to a million cycles for one complete inference. In the robust mode, the attack success rates on all four DNNs are more than 96.2% with an average glitch intensity of 1.4% and a maximum glitch intensity of 10.2%." @default.
• W3113404214 created "2021-01-05" @default.
• W3113404214 creator A5000387478 @default.
• W3113404214 creator A5029335324 @default.
• W3113404214 creator A5062851425 @default.
• W3113404214 date "2021-01-01" @default.
• W3113404214 modified "2023-09-26" @default.
• W3113404214 title "Stealthy and Robust Glitch Injection Attack on Deep Learning Accelerator for Target With Variational Viewpoint" @default.
• W3113404214 cites W1600075072 @default.
• W3113404214 cites W1849277567 @default.
• W3113404214 cites W2062272401 @default.
• W3113404214 cites W2125640276 @default.
• W3113404214 cites W2145862222 @default.
• W3113404214 cites W2194775991 @default.
• W3113404214 cites W2289252105 @default.
• W3113404214 cites W2302635400 @default.
• W3113404214 cites W2535873859 @default.
• W3113404214 cites W2543927648 @default.
• W3113404214 cites W2565960208 @default.
• W3113404214 cites W2604319603 @default.
• W3113404214 cites W2606722458 @default.
• W3113404214 cites W2793950911 @default.
• W3113404214 cites W2795915628 @default.
• W3113404214 cites W2798302089 @default.
• W3113404214 cites W2807835252 @default.
• W3113404214 cites W2887603965 @default.
• W3113404214 cites W2919115771 @default.
• W3113404214 cites W2943220429 @default.
• W3113404214 cites W2946801000 @default.
• W3113404214 cites W2950865323 @default.
• W3113404214 cites W2963163009 @default.
• W3113404214 cites W2963857521 @default.
• W3113404214 cites W2964097310 @default.
• W3113404214 cites W2973780393 @default.
• W3113404214 cites W2983058739 @default.
• W3113404214 cites W2985535619 @default.
• W3113404214 cites W2993396030 @default.
• W3113404214 cites W3008727730 @default.
• W3113404214 cites W3015806656 @default.
• W3113404214 cites W3092454700 @default.
• W3113404214 cites W3092516112 @default.
• W3113404214 cites W3105009650 @default.
• W3113404214 cites W4242053016 @default.
• W3113404214 cites W4288337628 @default.
• W3113404214 doi "https://doi.org/10.1109/tifs.2020.3046858" @default.
• W3113404214 hasPublicationYear "2021" @default.
• W3113404214 type Work @default.
• W3113404214 sameAs 3113404214 @default.
• W3113404214 citedByCount "4" @default.
• W3113404214 countsByYear W31134042142021 @default.
• W3113404214 countsByYear W31134042142022 @default.
• W3113404214 crossrefType "journal-article" @default.
• W3113404214 hasAuthorship W3113404214A5000387478 @default.
• W3113404214 hasAuthorship W3113404214A5029335324 @default.
• W3113404214 hasAuthorship W3113404214A5062851425 @default.
• W3113404214 hasBestOaLocation W31134042142 @default.
• W3113404214 hasConcept C108583219 @default.
• W3113404214 hasConcept C111919701 @default.
• W3113404214 hasConcept C113775141 @default.
• W3113404214 hasConcept C11413529 @default.
• W3113404214 hasConcept C138236772 @default.
• W3113404214 hasConcept C154945302 @default.
• W3113404214 hasConcept C162307627 @default.
• W3113404214 hasConcept C191287063 @default.
• W3113404214 hasConcept C199360897 @default.
• W3113404214 hasConcept C2777062904 @default.
• W3113404214 hasConcept C2777904410 @default.
• W3113404214 hasConcept C31972630 @default.
• W3113404214 hasConcept C41008148 @default.
• W3113404214 hasConcept C42935608 @default.
• W3113404214 hasConcept C50644808 @default.
• W3113404214 hasConcept C76155785 @default.
• W3113404214 hasConcept C79974875 @default.
• W3113404214 hasConcept C9390403 @default.
• W3113404214 hasConcept C94915269 @default.
• W3113404214 hasConceptScore W3113404214C108583219 @default.
• W3113404214 hasConceptScore W3113404214C111919701 @default.
• W3113404214 hasConceptScore W3113404214C113775141 @default.
• W3113404214 hasConceptScore W3113404214C11413529 @default.
• W3113404214 hasConceptScore W3113404214C138236772 @default.
• W3113404214 hasConceptScore W3113404214C154945302 @default.
• W3113404214 hasConceptScore W3113404214C162307627 @default.
• W3113404214 hasConceptScore W3113404214C191287063 @default.
• W3113404214 hasConceptScore W3113404214C199360897 @default.
• W3113404214 hasConceptScore W3113404214C2777062904 @default.
• W3113404214 hasConceptScore W3113404214C2777904410 @default.
• W3113404214 hasConceptScore W3113404214C31972630 @default.
• W3113404214 hasConceptScore W3113404214C41008148 @default.
• W3113404214 hasConceptScore W3113404214C42935608 @default.
• W3113404214 hasConceptScore W3113404214C50644808 @default.
• W3113404214 hasConceptScore W3113404214C76155785 @default.
• W3113404214 hasConceptScore W3113404214C79974875 @default.
• W3113404214 hasConceptScore W3113404214C9390403 @default.
• W3113404214 hasConceptScore W3113404214C94915269 @default.
• W3113404214 hasFunder F4320320709 @default.
• W3113404214 hasFunder F4320321001 @default.
• W3113404214 hasLocation W31134042141 @default.
• W3113404214 hasLocation W31134042142 @default.

• next