What is Diffusion Model?

Full Definition

Diffusion models are a class of generative models that learn to create data (typically images) by training on the task of reversing a noise-addition process. During training, Gaussian noise is iteratively added to real images across many steps until the image is pure noise; the model learns to predict and remove this noise at each step. At inference, the model starts from random noise and iteratively denoises it, guided by a text prompt or other conditioning signal, to produce a new image. Latent diffusion models (the technology behind Stable Diffusion, DALL-E 3, Midjourney, and Flux) operate in a compressed latent space rather than pixel space, making them far more efficient. Diffusion models have become the dominant paradigm for high-quality image generation, surpassing earlier GANs in diversity and controllability.