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 are susceptible to adversarial perturbations.Adversarial training and adversarial purification are among the most widelyrecognized defense strategies. Although these methods have different underlyinglogic, both rely on absolute logit values to generate label predictions. Inthis study, we theoretically analyze the logit difference around successfuladversarial attacks from a theoretical point of view and propose a newprinciple, namely Adversarial Logit Update (ALU), to infer adversarial sample'slabels. Based on ALU, we introduce a new classification paradigm that utilizespre- and post-purification logit differences for model's adversarial robustnessboost. Without requiring adversarial or additional data for model training, ourclean data synthesis model can be easily applied to various pre-trained modelsfor both adversarial sample detection and ALU-based data classification.Extensive experiments on both CIFAR-10, CIFAR-100, and tiny-ImageNet datasetsshow that even with simple components, the proposed solution achieves superiorrobustness performance compared to state-of-the-art methods against a widerange of adversarial attacks. Our python implementation is submitted in ourSupplementary document and will be published upon the paper's acceptance.

Advancing Adversarial Robustness Through Adversarial Logit Update