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

State-of-the-art deep learning models for tabular data have recently achievedacceptable performance to be deployed in industrial settings. However, therobustness of these models remains scarcely explored. Contrary to computervision, there is to date no realistic protocol to properly evaluate theadversarial robustness of deep tabular models due to intrinsic properties oftabular data such as categorical features, immutability, and featurerelationship constraints. To fill this gap, we propose CAA, the first efficientevasion attack for constrained tabular deep learning models. CAA is aniterative parameter-free attack that combines gradient and search attacks togenerate adversarial examples under constraints. We leverage CAA to build abenchmark of deep tabular models across three popular use cases: creditscoring, phishing and botnet attacks detection. Our benchmark supports tenthreat models with increasing capabilities of the attacker, and reflectsreal-world attack scenarios for each use case. Overall, our results demonstratehow domain knowledge, adversarial training, and attack budgets impact therobustness assessment of deep tabular models and provide security practitionerswith a set of recommendations to improve the robustness of deep tabular modelsagainst various evasion attack scenarios.

Constrained Adaptive Attacks: Realistic Evaluation of Adversarial Examples and Robust Training of Deep Neural Networks for Tabular Data