Goodfellow et al. demonstrate that adversarial misclassification works in the physical world: https://arxiv.org/abs/1607.02533 https://arxiv.org/abs/1607.02533

Goodfellow et al. demonstrate that adversarial misclassification works in the physical world: https://arxiv.org/abs/1607.02533 https://arxiv.org/abs/1607.02533 -- a printed image is consistently misclassified (demo here: https://www.youtube.com/watch?v=zQ_uMenoBCk https://www.youtube.com/watch?v=zQ_uMenoBCk) Recent work by me and some friends demonstrates that physical-world adversarial examples can actually be made quite robust, and you can synthesize 3D adversarial objects as well, and make them consistently classify as a desired target class: http://www.labsix.org/physical-objects-that-fool-neural-nets/ http://www.labsix.org/physical-objects-that-fool-neural-nets...