This document summarizes a lecture on generative adversarial networks (GANs). It introduces GANs as a type of implicit generative model that uses two neural networks (a generator and discriminator) competing against each other. The generator tries to generate realistic samples, while the discriminator tries to distinguish real and generated samples. Together they are trained to improve the generator's ability to generate high quality samples matching the real data distribution. Examples of GAN-generated images show their ability to produce sharp, realistic images.