M22.4 CONNECT THE MECHANISM
Train a generator against a learned discriminator
A forger learns by fooling a detective who keeps getting sharper. That game produced the first photorealistic fake faces, and a famous way of failing. Work out the detective's best strategy exactly.
LESSON OVERVIEW14 min lesson
Lesson overview
A forger learns by fooling a detective who keeps getting sharper. That game produced the first photorealistic fake faces, and a famous way of failing. Work out the detective's best strategy exactly.
What you’ll explore
- A GAN learns through competing generator and discriminator objectives; alternating updates can create useful samples but also instability, mode collapse, and misleading quality impressions.
GO TO THE SOURCE
Original explanations, connected to the research.
Generative Adversarial Networks (Goodfellow et al., 2014)Wasserstein GAN (Arjovsky, Chintala & Bottou, 2017)A Style-Based Generator Architecture for Generative Adversarial Networks (StyleGAN; Karras, Laine & Aila, 2018)Diffusion Models Beat GANs on Image Synthesis (Dhariwal & Nichol, 2021)Suggest a correction
A precise note can make an explanation better.
Choose the scene and describe what needs attention. Download a feedback file to share through a channel you already use. This page does not send feedback or connect you with a reviewer.