x
Attacking machine learning with adversarial examples — LessWrong