Adversarial ink: componentwise backward error attacks on deep learning

DOI10.1093/IMAMAT/HXAD017MaRDI QIDQ6630851zbMATH OpenFDO

Authors Lucas Beerens, Desmond J. Higham

Publication date 31 October 2024

Published in IMA Journal of Applied Mathematics (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://doi.org/10.1093/IMAMAT/HXAD017

zbMATH Keywords

conditioningmisclassificationoptimizationstability

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07)

Recommendations

• Adversarial noise attacks of deep learning architectures: stability analysis via sparse-modeled signals
• Analysis of classifiers' robustness to adversarial perturbations
• Detection of iterative adversarial attacks via counter attack
• Theoretical investigation of generalization bounds for adversarial learning of deep neural networks

Cited in

(3)

• Mappings, dimensionality and reversing out of deep neural networks
• The mathematics of adversarial attacks in AI -- why deep learning is unstable despite the existence of stable neural networks
• Do stable neural networks exist for classification problems? -- A new view on stability in AI

This page was built for publication: Adversarial ink: componentwise backward error attacks on deep learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6630851)