GAP++: Learning to generate target-conditioned adversarial examples
arXiv:2006.05097
Abstract
Adversarial examples are perturbed inputs which can cause a serious threat for machine learning models. Finding these perturbations is such a hard task that we can only use the iterative methods to traverse. For computational efficiency, recent works use adversarial generative networks to model the distribution of both the universal or image-dependent perturbations directly. However, these methods generate perturbations only rely on input images. In this work, we propose a more general-purpose framework which infers target-conditioned perturbations dependent on both input image and target label. Different from previous single-target attack models, our model can conduct target-conditioned attacks by learning the relations of attack target and the semantics in image. Using extensive experiments on the datasets of MNIST and CIFAR10, we show that our method achieves superior performance with single target attack models and obtains high fooling rates with small perturbation norms.
Accepted to IJCAI 2019 AIBS Workshop
References in corpus (9)
- Explaining and Harnessing Adversarial Examples
- Intriguing properties of neural networks
- Generating Natural Adversarial Examples
- Generating Adversarial Examples with Adversarial Networks
- Adversarial Transformation Networks: Learning to Generate Adversarial Examples
- Constructing Unrestricted Adversarial Examples with Generative Models
- Bilinear Representation for Language-based Image Editing Using Conditional Generative Adversarial Networks
- Regional Homogeneity: Towards Learning Transferable Universal Adversarial Perturbations Against Defenses
- Customizing an Adversarial Example Generator with Class-Conditional GANs