Matches in SemOpenAlex for { <https://semopenalex.org/work/W3080490461> ?p ?o ?g. }
Showing items 1 to 93 of
93
with 100 items per page.
- W3080490461 abstract "Deep Neural Networks (DNNs) have been shown to be prone to adversarial attacks. Memristive crossbars, being able to perform Matrix-Vector-Multiplications (MVMs) efficiently, are used to realize DNNs on hardware. However, crossbar non-idealities have always been devalued since they cause errors in performing MVMs, leading to computational accuracy losses in DNNs. Several software-based defenses have been proposed to make DNNs adversarially robust. However, no previous work has demonstrated the advantage conferred by the crossbar non-idealities in unleashing adversarial robustness. We show that the intrinsic hardware non-idealities yield adversarial robustness to the mapped DNNs without any additional optimization. We evaluate the adversarial resilience of state-of-the-art DNNs (VGG8 & VGG16 networks) using benchmark datasets (CIFAR-10, CIFAR-100 & Tiny Imagenet) across various crossbar sizes. We find that crossbar non-idealities unleash significantly greater adversarial robustness (>10-20%) in crossbar-mapped DNNs than baseline software DNNs. We further assess the performance of our approach with other state-of-the-art efficiency-driven adversarial defenses and find that our approach performs significantly well in terms of reducing adversarial loss." @default.
- W3080490461 created "2020-09-01" @default.
- W3080490461 creator A5004629816 @default.
- W3080490461 creator A5050310538 @default.
- W3080490461 date "2020-08-25" @default.
- W3080490461 modified "2023-10-11" @default.
- W3080490461 title "Rethinking Non-idealities in Memristive Crossbars for Adversarial Robustness in Neural Networks." @default.
- W3080490461 cites W1945616565 @default.
- W3080490461 cites W2082311137 @default.
- W3080490461 cites W2611005572 @default.
- W3080490461 cites W2620342231 @default.
- W3080490461 cites W2640329709 @default.
- W3080490461 cites W2909715355 @default.
- W3080490461 cites W2913848079 @default.
- W3080490461 cites W2951055820 @default.
- W3080490461 cites W2951807304 @default.
- W3080490461 cites W2954115294 @default.
- W3080490461 cites W2964014389 @default.
- W3080490461 cites W2997532515 @default.
- W3080490461 cites W3011221635 @default.
- W3080490461 cites W3017820626 @default.
- W3080490461 cites W3033076639 @default.
- W3080490461 cites W3106392217 @default.
- W3080490461 hasPublicationYear "2020" @default.
- W3080490461 type Work @default.
- W3080490461 sameAs 3080490461 @default.
- W3080490461 citedByCount "5" @default.
- W3080490461 countsByYear W30804904612020 @default.
- W3080490461 countsByYear W30804904612021 @default.
- W3080490461 crossrefType "posted-content" @default.
- W3080490461 hasAuthorship W3080490461A5004629816 @default.
- W3080490461 hasAuthorship W3080490461A5050310538 @default.
- W3080490461 hasConcept C104317684 @default.
- W3080490461 hasConcept C113775141 @default.
- W3080490461 hasConcept C13280743 @default.
- W3080490461 hasConcept C154945302 @default.
- W3080490461 hasConcept C185592680 @default.
- W3080490461 hasConcept C185798385 @default.
- W3080490461 hasConcept C199360897 @default.
- W3080490461 hasConcept C205649164 @default.
- W3080490461 hasConcept C2777904410 @default.
- W3080490461 hasConcept C2984842247 @default.
- W3080490461 hasConcept C29984679 @default.
- W3080490461 hasConcept C37736160 @default.
- W3080490461 hasConcept C41008148 @default.
- W3080490461 hasConcept C50644808 @default.
- W3080490461 hasConcept C55493867 @default.
- W3080490461 hasConcept C63479239 @default.
- W3080490461 hasConcept C76155785 @default.
- W3080490461 hasConceptScore W3080490461C104317684 @default.
- W3080490461 hasConceptScore W3080490461C113775141 @default.
- W3080490461 hasConceptScore W3080490461C13280743 @default.
- W3080490461 hasConceptScore W3080490461C154945302 @default.
- W3080490461 hasConceptScore W3080490461C185592680 @default.
- W3080490461 hasConceptScore W3080490461C185798385 @default.
- W3080490461 hasConceptScore W3080490461C199360897 @default.
- W3080490461 hasConceptScore W3080490461C205649164 @default.
- W3080490461 hasConceptScore W3080490461C2777904410 @default.
- W3080490461 hasConceptScore W3080490461C2984842247 @default.
- W3080490461 hasConceptScore W3080490461C29984679 @default.
- W3080490461 hasConceptScore W3080490461C37736160 @default.
- W3080490461 hasConceptScore W3080490461C41008148 @default.
- W3080490461 hasConceptScore W3080490461C50644808 @default.
- W3080490461 hasConceptScore W3080490461C55493867 @default.
- W3080490461 hasConceptScore W3080490461C63479239 @default.
- W3080490461 hasConceptScore W3080490461C76155785 @default.
- W3080490461 hasLocation W30804904611 @default.
- W3080490461 hasOpenAccess W3080490461 @default.
- W3080490461 hasPrimaryLocation W30804904611 @default.
- W3080490461 hasRelatedWork W2004823737 @default.
- W3080490461 hasRelatedWork W2503523779 @default.
- W3080490461 hasRelatedWork W2604505099 @default.
- W3080490461 hasRelatedWork W2605831051 @default.
- W3080490461 hasRelatedWork W2793518327 @default.
- W3080490461 hasRelatedWork W2807398060 @default.
- W3080490461 hasRelatedWork W2945824031 @default.
- W3080490461 hasRelatedWork W2951055820 @default.
- W3080490461 hasRelatedWork W2962700793 @default.
- W3080490461 hasRelatedWork W2963154688 @default.
- W3080490461 hasRelatedWork W2963207607 @default.
- W3080490461 hasRelatedWork W2963485691 @default.
- W3080490461 hasRelatedWork W2963752115 @default.
- W3080490461 hasRelatedWork W2964253222 @default.
- W3080490461 hasRelatedWork W2981639301 @default.
- W3080490461 hasRelatedWork W3014982797 @default.
- W3080490461 hasRelatedWork W3016235816 @default.
- W3080490461 hasRelatedWork W3017820626 @default.
- W3080490461 hasRelatedWork W3018820051 @default.
- W3080490461 hasRelatedWork W3189972411 @default.
- W3080490461 isParatext "false" @default.
- W3080490461 isRetracted "false" @default.
- W3080490461 magId "3080490461" @default.
- W3080490461 workType "article" @default.