Robust Adversarial Inputs Challenge Autonomous Vehicle Safety Claims
Original title: Robust adversarial inputs
45AI Score
Researchers created images that reliably fool neural network classifiers across varied scales and perspectives. This finding directly challenges recent claims that multi-view data collection makes autonomous vehicles resistant to malicious adversarial attacks.
Related Entities
Full Report
We’ve created images that reliably fool neural network classifiers when viewed from varied scales and perspectives. This challenges a claim from last week that self-driving cars would be hard to trick maliciously since they capture images from multiple scales, angles, perspectives, and the like.