ZK-GanDef: A GAN based Zero Knowledge Adversarial Training Defense for Neural Networks
This addresses the challenge of adaptive and efficient defense against adversarial attacks for neural network security, though it is incremental as it builds on existing GAN and adversarial training methods.
The paper tackles the problem of defending neural networks against adversarial examples by proposing ZK-GanDef, a GAN-based zero-knowledge adversarial training defense that does not require adversarial examples during training. The result shows it enhances test accuracy on adversarial examples by up to 49.17% compared to zero-knowledge approaches and achieves close performance to state-of-the-art full-knowledge methods with a maximum degradation of 8.46%, while being more efficient in training time.
Neural Network classifiers have been used successfully in a wide range of applications. However, their underlying assumption of attack free environment has been defied by adversarial examples. Researchers tried to develop defenses; however, existing approaches are still far from providing effective solutions to this evolving problem. In this paper, we design a generative adversarial net (GAN) based zero knowledge adversarial training defense, dubbed ZK-GanDef, which does not consume adversarial examples during training. Therefore, ZK-GanDef is not only efficient in training but also adaptive to new adversarial examples. This advantage comes at the cost of small degradation in test accuracy compared to full knowledge approaches. Our experiments show that ZK-GanDef enhances test accuracy on adversarial examples by up-to 49.17% compared to zero knowledge approaches. More importantly, its test accuracy is close to that of the state-of-the-art full knowledge approaches (maximum degradation of 8.46%), while taking much less training time.