In this talk I review recent work on three classes of models: implicit quantile networks, reverse-time diffusion models, and physics-informed neural networks. The mathematical underpinning of each of these models is reviewed and I illustrate the utility of these models using a disparate set of examples.