diffusion model

A diffusion model is a type of generative machine learning model that produces data samples (such as images, audio, or text) by simulating a gradual denoising process. It starts with random noise and iteratively refines it using learned patterns from training data, effectively reversing a “diffusion” process that successively added noise to data during training. Diffusion models have become popular for tasks like image synthesis because they can generate high-quality, realistic outputs.