CMI Seminar: Eliza O'Reilly
Determinantal point processes (DPPs) are a useful probabilistic model of point configurations exhibiting repulsion between points. They appear in random matrix theory and have found many applications including in machine learning for selecting diverse subsets. We will discuss the appealing properties of DPPs as well as a procedure for their simulation. I will then present a coupling result characterizing the repulsive effect of a point in a DPP, based on joint work with Jesper Møller.