Join or Log in
In a newly revised paper in Computer Vision and Pattern Recognition, a group of French and Swiss computer science researchers show that "a very small perturbation vector that causes natural images to be misclassified with high probability" -- that is, a minor image transformation can beat machine learning systems nearly every time.
The "universal adversarial preturbation" undetectably alters images so AI can't recognize them 
Added 2 years ago by Francis Tseng
Show info
The "universal adversarial preturbation" undetectably alters images so AI can't recognize them 
Info
In a newly revised paper in Computer Vision and Pattern Recognition, a group of French and Swiss computer science researchers show that "a very small perturbation vector that causes natural images to be misclassified with high probability" -- that is, a minor image transformation can beat machine learning systems nearly every time.
1 Connection