Robustness should not be at odds with accuracy

DOI10.4230/LIPICS.FORC.2022.5MaRDI QIDQ6128556zbMATH OpenFDO

Authors Ruth Urner

Publication date 15 April 2024

Full work available at URL https://doi.org/10.4230/LIPICS.FORC.2022.5

zbMATH Keywords

adaptive robust lossadversarial perturbationsBayes optimal classifierstatistical learning theory

Mathematics Subject Classification ID

Privacy of data (68P27) Artificial intelligence (68Txx)

Recommendations

• Achieving adversarial robustness requires an active teacher
• Analysis of classifiers' robustness to adversarial perturbations
• Precise statistical analysis of classification accuracies for adversarial training
• When Should You Defend Your Classifier?
• Adversarial robustness for latent models: revisiting the robust-standard accuracies tradeoff

Cited in

(3)

• Precise statistical analysis of classification accuracies for adversarial training
• Adversarial robustness for latent models: revisiting the robust-standard accuracies tradeoff
• Improving the accuracy-robustness trade-off of classifiers via adaptive smoothing

This page was built for publication: Robustness should not be at odds with accuracy

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6128556)