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) are often criticized for being susceptible toadversarial attacks. Most successful defense strategies adopt adversarialtraining or random input transformations that typically require retraining orfine-tuning the model to achieve reasonable performance. In this work, ourinvestigations of intermediate representations of a pre-trained DNN lead to aninteresting discovery pointing to intrinsic robustness to adversarial attacks.We find that we can learn a generative classifier by statisticallycharacterizing the neural response of an intermediate layer to clean trainingsamples. The predictions of multiple such intermediate-layer based classifiers,when aggregated, show unexpected robustness to adversarial attacks.Specifically, we devise an ensemble of these generative classifiers thatrank-aggregates their predictions via a Borda count-based consensus. Ourproposed approach uses a subset of the clean training data and a pre-trainedmodel, and yet is agnostic to network architectures or the adversarial attackgeneration method. We show extensive experiments to establish that our defensestrategy achieves state-of-the-art performance on the ImageNet validation set.

REGroup: Rank-aggregating Ensemble of Generative Classifiers for Robust Predictions