New Method Exposes How Artificial Intelligence Works

The neural networks are harder to fool thanks to adversarial training. Los Alamos National Laboratory researchers have developed a novel method for comparing neural networks...