Many important semi-supervised machine learning problems arise in complex geometric environments. It is natural that there is a wide variety of different geometries, and the best performing learning algorithm varies depending on the application. One of the objectives in this project is to develop rather general theoretical foundations using stochastic game theory, partial differential equations and geometric analysis in a way that would also apply to graph-based semi-supervised learning problems. The problems we propose are interesting theoretical open problems in their respective fields as well.

2024

2028