Adversarial Attack Beats Near-Superhuman AI Go Player

Computer security specialists had hoped Go-playing AI agents would be immune to adversarial attacks. Now it’s back to the drawing board. Read more

Similar