Unifying Adversarial Training Algorithms with Flexible Deep Data Gradient Regularization

Abstract

Many previous proposals for adversarial training of deep neural nets have included di- rectly modifying the gradient, training on a mix of original and adversarial examples, using contractive penalties, and approximately optimizing constrained adversarial ob- jective functions. In this paper, we show these proposals are actually all instances of optimizing a general, regularized objective we call DataGrad. Our proposed DataGrad framework, which can be viewed as a deep extension of the layerwise contractive au- toencoder penalty, cleanly simplifies prior work and easily allows extensions such as adversarial training with multi-task cues. In our experiments, we find that the deep gra- dient regularization of DataGrad (which also has L1 and L2 flavors of regularization) outperforms alternative forms of regularization, including classical L1, L2, and multi- task, both on the original dataset as well as on adversarial sets. Furthermore, we find that combining multi-task optimization with DataGrad adversarial training results in the most robust performance.

References in corpus (5)

• Explaining and Harnessing Adversarial Examples
• Towards Deep Neural Network Architectures Robust to Adversarial Examples
• Learning with a Strong Adversary
• A Unified Gradient Regularization Family for Adversarial Examples
• Improving Back-Propagation by Adding an Adversarial Gradient

Cited by in corpus (7)

• A trans-disciplinary review of deep learning research for water resources scientists
• Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey
• Robust Learning with Jacobian Regularization
• Learning Adversary-Resistant Deep Neural Networks
• Interpreting Adversarial Robustness: A View from Decision Surface in Input Space
• Using Non-invertible Data Transformations to Build Adversarial-Robust Neural Networks
• Adversary Resistant Deep Neural Networks with an Application to Malware Detection