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

Compared to human vision, computer vision based on convolutional neuralnetworks (CNN) are more vulnerable to adversarial noises. This difference islikely attributable to how the eyes sample visual input and how the brainprocesses retinal samples through its dorsal and ventral visual pathways, whichare under-explored for computer vision. Inspired by the brain, we designrecurrent neural networks, including an input sampler that mimics the humanretina, a dorsal network that guides where to look next, and a ventral networkthat represents the retinal samples. Taking these modules together, the modelslearn to take multiple glances at an image, attend to a salient part at eachglance, and accumulate the representation over time to recognize the image. Wetest such models for their robustness against a varying level of adversarialnoises with a special focus on the effect of different input samplingstrategies. Our findings suggest that retinal foveation and sampling renders amodel more robust against adversarial noises, and the model may correct itselffrom an attack when it is given a longer time to take more glances at an image.In conclusion, robust visual recognition can benefit from the combined use ofthree brain-inspired mechanisms: retinal transformation, attention guided eyemovement, and recurrent processing, as opposed to feedforward-only CNNs.

Human Eyes Inspired Recurrent Neural Networks are More Robust Against Adversarial Noises