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

Recently it has been shown that state-of-the-art NLP models are vulnerable toadversarial attacks, where the predictions of a model can be drasticallyaltered by slight modifications to the input (such as synonym substitutions).While several defense techniques have been proposed, and adapted, to thediscrete nature of text adversarial attacks, the benefits of general-purposeregularization methods such as label smoothing for language models, have notbeen studied. In this paper, we study the adversarial robustness provided byvarious label smoothing strategies in foundational models for diverse NLP tasksin both in-domain and out-of-domain settings. Our experiments show that labelsmoothing significantly improves adversarial robustness in pre-trained modelslike BERT, against various popular attacks. We also analyze the relationshipbetween prediction confidence and robustness, showing that label smoothingreduces over-confident errors on adversarial examples.

In and Out-of-Domain Text Adversarial Robustness via Label Smoothing