From the course: Digital Marketing Trends

Unlock the full course today

Join today to access over 22,400 courses taught by industry experts or purchase this course individually.

GANs: Generative adversarial networks

GANs: Generative adversarial networks

From the course: Digital Marketing Trends

Start my 1-month free trial

GANs: Generative adversarial networks

In simple terms, they're a bit like GANs are made up of two parts. like images, text, audio and videos from the data it's fed The discriminator, which is fed the same data and show why it's output isn't up to par and the generator takes that learning, tries again and again until it develops an output where the discriminator can give it the AI version of a thumbs up. but GANs learn through challenge and conflict and the results they produce are often surprising. They're responsible for creating some pretty disturbing AI art. GANs are also the reason deepfake videos look so realistic. Deepfakes happen when the image of one person is put onto the face and body of another. They're like a super realistic face swap and the results can be so believable, it's hard to tell if they're fake. Research company, Open AI, created a natural language processor that uses GANs to write pretty convincing copy and like deepfake videos, it's difficult to distinguish what's human and real and what's not. So…

Contents