AlgoPlus//machine learning / gan
Read the theory

Generative Adversarial Network

A generator and a discriminator train against each other until the fakes look real.

Round
Discriminator
Legend
Real data
Generated
AI Tutor Workspace
In a nutshell
A GAN trains two neural networks against each other. The generator turns random noise into fake samples; the discriminator tries to tell those fakes apart from real training data. Each round the discriminator gets better at catching fakes, which forces the generator to make them more convincing — until the fakes are so realistic the discriminator can only guess (about 50/50), meaning the generator has learned to reproduce the real data.
Ready
Press play to begin the cinematic walkthrough.
A forger and a detective locked in a duel. Every time the detective spots a fake, the forger improves — until the counterfeits are indistinguishable and the detective can only guess.
Key terms
Go deeper in the lesson
Read the full theory, intuition & complexity for Generative Adversarial Network.