GanDef: A GAN based Adversarial Training Defense for Neural Network Classifier

Abstract

Machine learning models, especially neural network (NN) classifiers, are widely used in many applications including natural language processing, computer vision and cybersecurity. They provide high accuracy under the assumption of attack-free scenarios. However, this assumption has been defied by the introduction of adversarial examples -- carefully perturbed samples of input that are usually misclassified. Many researchers have tried to develop a defense against adversarial examples; however, we are still far from achieving that goal. In this paper, we design a Generative Adversarial Net (GAN) based adversarial training defense, dubbed GanDef, which utilizes a competition game to regulate the feature selection during the training. We analytically show that GanDef can train a classifier so it can defend against adversarial examples. Through extensive evaluation on different white-box adversarial examples, the classifier trained by GanDef shows the same level of test accuracy as those trained by state-of-the-art adversarial training defenses. More importantly, GanDef-Comb, a variant of GanDef, could utilize the discriminator to achieve a dynamic trade-off between correctly classifying original and adversarial examples. As a result, it achieves the highest overall test accuracy when the ratio of adversarial examples exceeds 41.7%.

References in corpus (6)

• Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
• Technical Report on the CleverHans v2.1.0 Adversarial Examples Library
• Defense-GAN: Protecting Classifiers Against Adversarial Attacks Using Generative Models
• Learning to Pivot with Adversarial Networks
• Extending Defensive Distillation
• Fortified Networks: Improving the Robustness of Deep Networks by Modeling the Manifold of Hidden Representations

Cited by in corpus (3)

• Generative Adversarial Networks: A Survey Towards Private and Secure Applications
• Trust but Verify: An Information-Theoretic Explanation for the Adversarial Fragility of Machine Learning Systems, and a General Defense against Adversarial Attacks
• ManiGen: A Manifold Aided Black-box Generator of Adversarial Examples