Generate (non-software) Bugs to Fool Classifiers
This addresses security vulnerabilities in autonomous systems like cars by creating stealthy attacks that humans might overlook, representing a novel but incremental advance in adversarial machine learning.
The paper tackles the problem of generating adversarial examples that appear natural to humans, such as insect images on stop signs or bird chirps in audio, to fool classification models, and demonstrates feasibility with generative adversarial networks and an optimization algorithm for fast, successful attacks in both image and audio domains.
In adversarial attacks intended to confound deep learning models, most studies have focused on limiting the magnitude of the modification so that humans do not notice the attack. On the other hand, during an attack against autonomous cars, for example, most drivers would not find it strange if a small insect image were placed on a stop sign, or they may overlook it. In this paper, we present a systematic approach to generate natural adversarial examples against classification models by employing such natural-appearing perturbations that imitate a certain object or signal. We first show the feasibility of this approach in an attack against an image classifier by employing generative adversarial networks that produce image patches that have the appearance of a natural object to fool the target model. We also introduce an algorithm to optimize placement of the perturbation in accordance with the input image, which makes the generation of adversarial examples fast and likely to succeed. Moreover, we experimentally show that the proposed approach can be extended to the audio domain, for example, to generate perturbations that sound like the chirping of birds to fool a speech classifier.