What you’re seeing in this video from graphics processing firm Nvidia is the result of two algorithmic adversaries trying to one-up each other. Working from a photo database of 30,000 celebrity faces, the two algorithms learned about different details, like beards and jewelry, that make a face look real to the human eye, and then engaged in a rapid-fire back-and-forth process that produced amazingly realistic results. None of the good-looking folks you see are real, but you’d never know it.
The algorithms are joined in what’s known as a GAN, or generative adversarial network, the idea being that two “heads” are better that one. The first one, the generator, creates a new composite face from the real celebrity data set, and the second net, the discriminator, tries to tell if it’s real or fake. Repeating this process causes the generator’s images to get better and better.
These generators can also work with photos of objects. GANs are a hot research area right now at companies like Google, Adobe, and Facebook, and one company is already marketing virtual fashion models to clothing retailers.