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

Despite high accuracy of Convolutional Neural Networks (CNNs), they arevulnerable to adversarial and out-distribution examples. There are manyproposed methods that tend to detect or make CNNs robust against these foolingexamples. However, most such methods need access to a wide range of foolingexamples to retrain the network or to tune detection parameters. Here, wepropose a method to detect adversarial and out-distribution examples against apre-trained CNN without needing to retrain the CNN or needing access to a widevariety of fooling examples. To this end, we create adversarial profiles foreach class using only one adversarial attack generation technique. We then wrapa detector around the pre-trained CNN that applies the created adversarialprofile to each input and uses the output to decide whether or not the input islegitimate. Our initial evaluation of this approach using MNIST dataset showthat adversarial profile based detection is effective in detecting at least 92of out-distribution examples and 59% of adversarial examples.

Adversarial Profiles: Detecting Out-Distribution & Adversarial Samples in Pre-trained CNNs