Mimic and Fool: A Task Agnostic Adversarial Attack
This addresses the need for more general adversarial attacks in computer vision, though it is incremental as it builds on existing feature extractor-based methods.
The paper tackles the problem of designing adversarial attacks that are task-agnostic by proposing Mimic and Fool, which finds adversarial images that mimic the features of original images to fool downstream computer vision tasks. The attack achieved success rates of 74.0%, 81.0%, and 87.1% on image captioning and VQA models.
At present, adversarial attacks are designed in a task-specific fashion. However, for downstream computer vision tasks such as image captioning, image segmentation etc., the current deep learning systems use an image classifier like VGG16, ResNet50, Inception-v3 etc. as a feature extractor. Keeping this in mind, we propose Mimic and Fool, a task agnostic adversarial attack. Given a feature extractor, the proposed attack finds an adversarial image which can mimic the image feature of the original image. This ensures that the two images give the same (or similar) output regardless of the task. We randomly select 1000 MSCOCO validation images for experimentation. We perform experiments on two image captioning models, Show and Tell, Show Attend and Tell and one VQA model, namely, end-to-end neural module network (N2NMN). The proposed attack achieves success rate of 74.0%, 81.0% and 87.1% for Show and Tell, Show Attend and Tell and N2NMN respectively. We also propose a slight modification to our attack to generate natural-looking adversarial images. In addition, we also show the applicability of the proposed attack for invertible architecture. Since Mimic and Fool only requires information about the feature extractor of the model, it can be considered as a gray-box attack.