Robust Physical-World Attacks on Deep Learning Visual ...
This paper appears at CVPR 2018Robust Physical-World Attacks on Deep Learning Visual ClassificationKevin Eykholt 1, Ivan Evtimov*2, Earlence Fernandes2, Bo Li3,Amir Rahmati4, Chaowei Xiao1, Atul Prakash1, Tadayoshi Kohno2, and Dawn Song31University of Michigan, Ann Arbor2University of Washington3University of California, Berkeley4Samsung Research America and Stony Brook UniversityAbstractRecent studies show that the state-of-the-art deep neuralnetworks (DNNs) are vulnerable to adversarial examples,resulting from small-magnitude perturbations added to theinput.
(2) Road signs exist in a noisy unconstrained environment with changing physical conditions such as the distance and angle of the viewing camera, implying that physical adversarial perturba-tions should be robust against considerable environmental instability. (3) Road signs play an important role in trans-portation safety.
Download Robust Physical-World Attacks on Deep Learning Visual ...
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: