CASUS Institute Seminar, Ismael Medina, Göttingen University (Germany)
Ismael is a doctoral student in the Optimal Transport Group led by Prof. Bernhard Schmitzer at the University’s Institute of Computer Science.
Optimal transport (OT) is the problem of moving the mass of a probability measure μ to match another probability measure ν in an optimal way with respect to some cost. This problem induces a distance in the space of probability measures with a wide range of applications, both within math (partial differential equations, optimization, statistics, etc.) as well as in other disciplines (image processing, data science, economics, biology, etc.).
In this talk Ismael will first lay the foundations of optimal transport. Then some of the
numerical methods for solving the OT problem will be discussed, including the domain decomposition algorithm for (entropic) OT, a fast and parallelizable method for big OT problems. Last, Ismael will review some of the applications of OT to other disciplines.