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

As humans, we inherently perceive images based on their predominant features,and ignore noise embedded within lower bit planes. On the contrary, Deep NeuralNetworks are known to confidently misclassify images corrupted withmeticulously crafted perturbations that are nearly imperceptible to the humaneye. In this work, we attempt to address this problem by training networks toform coarse impressions based on the information in higher bit planes, and usethe lower bit planes only to refine their prediction. We demonstrate that, byimposing consistency on the representations learned across differentlyquantized images, the adversarial robustness of networks improves significantlywhen compared to a normally trained model. Present state-of-the-art defensesagainst adversarial attacks require the networks to be explicitly trained usingadversarial samples that are computationally expensive to generate. While suchmethods that use adversarial training continue to achieve the best results,this work paves the way towards achieving robustness without having toexplicitly train on adversarial samples. The proposed approach is thereforefaster, and also closer to the natural learning process in humans.

Towards Achieving Adversarial Robustness by Enforcing Feature Consistency Across Bit Planes