The next big threat to AI might already be lurking on the web | ZDNET

"While large deep learning models are resilient to random noise, even minuscule amounts of adversarial noise in training sets (i.e., a poisoning …

Link to Full Article: Read Here