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

In practical cloud-edge scenarios, where a resource constrained edge performsdata acquisition and a cloud system (having sufficient resources) performsinference tasks with a deep neural network (DNN), adversarial robustness iscritical for reliability and ubiquitous deployment. Adversarial detection is aprime adversarial defence technique used in prior literature. However, in priordetection works, the detector is attached to the classifier model and bothdetector and classifier work in tandem to perform adversarial detection thatrequires a high computational overhead which is not available at the low-poweredge. Therefore, prior works can only perform adversarial detection at thecloud and not at the edge. This means that in case of adversarial attacks, theunfavourable adversarial samples must be communicated to the cloud which leadsto energy wastage at the edge device. Therefore, a low-power edge-friendlyadversarial detection method is required to improve the energy efficiency ofthe edge and robustness of the cloud-based classifier. To this end, RobustEdgeproposes Quantization-enabled Energy Separation (QES) training with "earlydetection and exit" to perform edge-based low cost adversarial detection. TheQES-trained detector implemented at the edge blocks adversarial datatransmission to the classifier model, thereby improving adversarial robustnessand energy-efficiency of the Cloud-Edge system.

RobustEdge: Low Power Adversarial Detection for Cloud-Edge Systems