An invitation to K\"ahler-Einstein metrics and random point processes

Abstract

This is an invitation to the probabilistic approach for constructing Kahler-Einstein metrics on complex projective algebraic manifolds X. The metrics in question emerge in the large N-limit from a canonical way of sampling N points on X, i.e. from random point processes on X, defined in terms of algebro-geometric data. The proof of the convergence towards Kahler-Einstein metrics with negative Ricci curvature is explained. In the case of positive Ricci curvature a variational approach is introduced to prove the conjectural convergence, which can be viewed as a probabilistic constructive analog of the Yau-Tian-Donaldson conjecture. The variational approach reveals, in particular, that the convergence holds under the hypothesis that there is no phase transition, which - from the algebro-geometric point of view - amounts to an analytic property of a certain Archimedean zeta function.