Matches in SemOpenAlex for { <https://semopenalex.org/work/W3183458034> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W3183458034 abstract "With a growing need to enable intelligence in embedded devices in the Internet of Things (IoT) era, secure hardware implementation of Deep Neural Networks (DNNs) has become imperative. We will focus on how to address adversarial robustness for DNNs through efficiency-driven hardware optimizations. Since memory (specifically, dot-product operations) is a key energy-spending component for DNNs, hardware approaches in the past have focused on optimizing the memory. One such approach is approximate digital CMOS memories with hybrid 6T-8T SRAM cells that enable supply voltage (Vdd) scaling yielding low-power operation, without significantly affecting the performance due to read/write failures incurred in the 6T cells. In this paper, we show how the bit-errors in the 6T cells of hybrid 6T-8T memories minimize the adversarial perturbations in a DNN. Essentially, we find that for different configurations of 8T-6T ratios and scaledVdd operation, noise incurred in the hybrid memory architectures is bound within specific limits. This hardware noise can potentially interfere in the creation of adversarial attacks in DNNs yielding robustness. Another memory optimization approach involves using analog memristive crossbars that perform Matrix-Vector-Multiplications (MVMs) efficiently with low energy and area requirements. However, crossbars generally suffer from intrinsic non-idealities that cause errors in performing MVMs, leading to degradation in the accuracy of the DNNs. We will show how the intrinsic hardware variations manifested through crossbar non-idealities yield adversarial robustness to the mapped DNNs without any additional optimization." @default.
- W3183458034 created "2021-08-02" @default.
- W3183458034 creator A5004629816 @default.
- W3183458034 creator A5050310538 @default.
- W3183458034 creator A5050796355 @default.
- W3183458034 date "2021-02-01" @default.
- W3183458034 modified "2023-10-16" @default.
- W3183458034 title "Efficiency-driven Hardware Optimization for Adversarially Robust Neural Networks" @default.
- W3183458034 cites W2004823737 @default.
- W3183458034 cites W2012194400 @default.
- W3183458034 cites W2063927304 @default.
- W3183458034 cites W2132621842 @default.
- W3183458034 cites W2166230011 @default.
- W3183458034 cites W2171894735 @default.
- W3183458034 cites W2289567715 @default.
- W3183458034 cites W2334564994 @default.
- W3183458034 cites W2603766943 @default.
- W3183458034 cites W2604505099 @default.
- W3183458034 cites W2951055820 @default.
- W3183458034 cites W2962700793 @default.
- W3183458034 cites W2963485691 @default.
- W3183458034 cites W2983353765 @default.
- W3183458034 cites W3033519076 @default.
- W3183458034 cites W3092585568 @default.
- W3183458034 cites W3106392217 @default.
- W3183458034 doi "https://doi.org/10.23919/date51398.2021.9474001" @default.
- W3183458034 hasPublicationYear "2021" @default.
- W3183458034 type Work @default.
- W3183458034 sameAs 3183458034 @default.
- W3183458034 citedByCount "6" @default.
- W3183458034 countsByYear W31834580342020 @default.
- W3183458034 countsByYear W31834580342022 @default.
- W3183458034 countsByYear W31834580342023 @default.
- W3183458034 crossrefType "proceedings-article" @default.
- W3183458034 hasAuthorship W3183458034A5004629816 @default.
- W3183458034 hasAuthorship W3183458034A5050310538 @default.
- W3183458034 hasAuthorship W3183458034A5050796355 @default.
- W3183458034 hasBestOaLocation W31834580342 @default.
- W3183458034 hasConcept C104317684 @default.
- W3183458034 hasConcept C113775141 @default.
- W3183458034 hasConcept C119599485 @default.
- W3183458034 hasConcept C127413603 @default.
- W3183458034 hasConcept C185592680 @default.
- W3183458034 hasConcept C24326235 @default.
- W3183458034 hasConcept C2742236 @default.
- W3183458034 hasConcept C41008148 @default.
- W3183458034 hasConcept C46362747 @default.
- W3183458034 hasConcept C55493867 @default.
- W3183458034 hasConcept C63479239 @default.
- W3183458034 hasConcept C68043766 @default.
- W3183458034 hasConcept C9390403 @default.
- W3183458034 hasConceptScore W3183458034C104317684 @default.
- W3183458034 hasConceptScore W3183458034C113775141 @default.
- W3183458034 hasConceptScore W3183458034C119599485 @default.
- W3183458034 hasConceptScore W3183458034C127413603 @default.
- W3183458034 hasConceptScore W3183458034C185592680 @default.
- W3183458034 hasConceptScore W3183458034C24326235 @default.
- W3183458034 hasConceptScore W3183458034C2742236 @default.
- W3183458034 hasConceptScore W3183458034C41008148 @default.
- W3183458034 hasConceptScore W3183458034C46362747 @default.
- W3183458034 hasConceptScore W3183458034C55493867 @default.
- W3183458034 hasConceptScore W3183458034C63479239 @default.
- W3183458034 hasConceptScore W3183458034C68043766 @default.
- W3183458034 hasConceptScore W3183458034C9390403 @default.
- W3183458034 hasFunder F4320306076 @default.
- W3183458034 hasLocation W31834580341 @default.
- W3183458034 hasLocation W31834580342 @default.
- W3183458034 hasOpenAccess W3183458034 @default.
- W3183458034 hasPrimaryLocation W31834580341 @default.
- W3183458034 hasRelatedWork W2012045996 @default.
- W3183458034 hasRelatedWork W2109451123 @default.
- W3183458034 hasRelatedWork W2967161359 @default.
- W3183458034 hasRelatedWork W3024050170 @default.
- W3183458034 hasRelatedWork W3151633427 @default.
- W3183458034 hasRelatedWork W3211992815 @default.
- W3183458034 hasRelatedWork W4210772589 @default.
- W3183458034 hasRelatedWork W4293253840 @default.
- W3183458034 hasRelatedWork W4308090481 @default.
- W3183458034 hasRelatedWork W4378977321 @default.
- W3183458034 isParatext "false" @default.
- W3183458034 isRetracted "false" @default.
- W3183458034 magId "3183458034" @default.
- W3183458034 workType "article" @default.