Adversarial defense via the data-dependent activation, total variation minimization, and adversarial training

DOI10.3934/ipi.2020046OpenAlexW3048203754MaRDI QIDQ2028930

Wei Zhu, Bao Wang, Penghang Yin, Stanley J. Osher, Andrea L. Bertozzi, Alex Tong Lin

Publication date: 3 June 2021

Published in: Inverse Problems and Imaging (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1809.08516

zbMATH Keywords

interpolationdeep learningadversarial defense

Mathematics Subject Classification ID

Machine vision and scene understanding (68T45) General topics in artificial intelligence (68T01)

Related Items (4)

EnResNet: ResNets Ensemble via the Feynman--Kac Formalism for Adversarial Defense and Beyond ⋮ Generalization Error Analysis of Neural Networks with Gradient Based Regularization ⋮ How does momentum benefit deep neural networks architecture design? A few case studies ⋮ Graph interpolating activation improves both natural and robust accuracies in data-efficient deep learning

Uses Software

• Adam

Cites Work

• Nonlinear total variation based noise removal algorithms

This page was built for publication: Adversarial defense via the data-dependent activation, total variation minimization, and adversarial training