English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep Neural Networks (DNNs) have been shown to be prone to adversarialattacks. Memristive crossbars, being able to performMatrix-Vector-Multiplications (MVMs) efficiently, are used to realize DNNs onhardware. However, crossbar non-idealities have always been devalued since theycause errors in performing MVMs, leading to computational accuracy losses inDNNs. Several software-based defenses have been proposed to make DNNsadversarially robust. However, no previous work has demonstrated the advantageconferred by the crossbar non-idealities in unleashing adversarial robustness.We show that the intrinsic hardware non-idealities yield adversarial robustnessto the mapped DNNs without any additional optimization. We evaluate theadversarial resilience of state-of-the-art DNNs (VGG8 & VGG16 networks) usingbenchmark datasets (CIFAR-10, CIFAR-100 & Tiny Imagenet) across variouscrossbar sizes. We find that crossbar non-idealities unleash significantlygreater adversarial robustness (>10-20%) in crossbar-mapped DNNs than baselinesoftware DNNs. We further assess the performance of our approach with otherstate-of-the-art efficiency-driven adversarial defenses and find that ourapproach performs significantly well in terms of reducing adversarial loss.

Rethinking Non-idealities in Memristive Crossbars for Adversarial Robustness in Neural Networks