Notes on generative modeling: flow matching, diffusion, optimal transport and Schr{ö}dinger bridge

The authors explain the main mathematical ideas behind several methods used to create generative models, which are techniques to generate new data similar to a given dataset. They highlight how optimal transport, a way to match different data distributions efficiently, relates to existing approaches like the Schrödinger bridge and flow matching methods. Their notes help connect these different concepts within generative modeling.