Diffusion models, also known as diffusion probabilistic models or score-based generative models, are a class of latent variable generative models used in machine learning. They consist of three major components: the forward process, the reverse process, and the sampling procedure.
One example of a diffusion model is the Denoising Autoencoder (DAE), which is trained to reconstruct a clean image from a corrupted version of the same image. The DAE consists of an encoder network that maps the corrupted image to a lower-dimensional latent space, and a decoder network that maps the latent space back to the original image.