Diffusion Models: A Comprehensive Survey of Methods and Applications

efficient sampling, improved likelihood estimation, and handling data with special structures.

shown potential in a variety of domains:

• computer vision
• natural language processing
• temporal data modeling
• multi-modal modeling
• robust machine learning
• interdisciplinary applications

diffusion models:

• denoising diffusion probabilistic models (DDPMs)
• score-based generative models (SGMs)
• stochastic differential equations (Score SDEs)

progressively perturb data with intensifying random noise (called the “diffusion” process)

then successively remove noise to generate new data samples