[Google Cloud Skills Boost(Qwiklabs)] Introduction to Generative AI Learning Path - 5. Introduction to Image Generation

https://www.youtube.com/playlist?list=PLIivdWyY5sqIlLF9JHbyiqzZbib9pFt4x

Generative AI Learning Path

---

• GANs 
  • One neural network (Generator) -> creates images
  • The other neural network (Discriminator) -> predicts if the image is real or fake

We add noise to an image iteratively and then train a model that learns how to de-noise an image.

• Forward diffusion
  • Start with a clean image and add noise iteratively
• Reverse diffusion
  • Start with a noisy image and remove noise iteratively

• From the pure noise, we could have a model that will be able to synthesize a novel image.

---