PopSkipJump: Decision-Based Attack for Probabilistic Classifiers
Carl-Johann Simon-Gabriel 1 Noman Ahmed Sheikh 1 Andreas Krause 1
Abstract noisy or probabilistic classification outputs – a quite natural
and common setting in the real world. There could indeed
Most current classifiers are vulnerable to adver-
be many reasons for this noise: occasional measurement
sarial examples, ...


雷达卡




京公网安备 11010802022788号







