ABSTRACT

In this talk, I will provide a bird’s-eye view on the mathematical foundations and related problems of training diffusion models in the context of generative AI. The main tools are from stochastic processes, control theory and partial differential equations.