Ein Diffusionsmodell ist ein generatives Modell, das darauf trainiert ist, einen Prozess umzukehren, der Daten schrittweise Rauschen hinzufügt. Indem es lernt, von zufälligem Rauschen zu einer Ziel-Datenverteilung hin zu entrauschen, kann es realistische Bilder, Audio, Textrepräsentationen oder andere Ausgaben erzeugen.
In machine learning, diffusion models, also known as diffusion probabilistic models or score-based generative models, are a class of latent variable models. They are Markov chains trained using variational inference. The goal of diffusion models is to learn the latent structure of a dataset by modeling the way in which data points diffuse through the latent space. In computer vision, this means that a neural network is trained to denoise images blurred with Gaussian noise by learning to reverse the diffusion process. It mainly consists of three major components: the forward process, the revers