Main Concept
A GAN is a generative model made up of two competing neural networks: a Generator (creates fake data) and a Discriminator (tries to distinguish fake from real). They train against each other in an adversarial process.
How It Works
- Generator — takes random noise and creates synthetic data (images, etc.)
- Discriminator — receives both real and fake data, tries to classify which is which
- Competition — as the discriminator gets better at detecting fakes, the generator gets better at fooling it
- Result — eventually, the generator produces data indistinguishable from the real thing
Use Cases
- Image generation — creating realistic artificial images
- Data augmentation — generating synthetic training data
- Style transfer — applying artistic styles to images
- Image-to-image translation — converting sketches to photos, day to night, etc.
- Super-resolution — upscaling low-resolution images
GAN vs. Diffusion Models
| GAN | Diffusion | |
|---|---|---|
| Training | Adversarial | Noise removal |
| Stability | Can be unstable | More stable |
| Quality | High quality but variable | Consistently high quality |
| Current use | Less common now | Preferred in modern GenAI |
Modern models like DALL-E, Stable Diffusion, and Amazon Nova Canvas use diffusion, not GANs, because they’re more stable and produce better results.
AIF-C01 Context
Know GANs as a generative AI approach for creating synthetic data. The exam may ask about methods for data augmentation or synthetic data generation — GANs are a classic answer, though diffusion models are more relevant today.