The Discrete Gaussian model is a type of integer-valued random height function. In the 2D setting, it exhibits a phase transition between a localised phase and a delocalised phase. This phenomenon is also called the Kosterlitz-Thouless phase transition, whose terminology originates from its dual counterpart, the planar XY model. Motivation for studying the Discrete Gaussian model is multifold. Due to its duality relations with a number of 2D mathematical physics models, such as the XY model or the Coulomb gas, studies on integer-value height functions are capable of proving a number of conjectures usually not accessible using classical methods. Other discrete height functions also have dualities with a number of different interesting models, so it will be of vast interest to develop a general framework that deals with discrete height functions. Also, discrete height functions are considered to be appropriate test cases for recently developed techniques from probability theory. In this talk, we discuss a particular method called the renormalisation group method, which is believed to serve as a general framework for studying random fields. We also discuss briefly how the renormalisation group method can be used to prove that the scaling limit of the 2D Discrete Gaussian model is a 2D Gaussian free field.

In this talk we shall first review our recent results about the equivalence of non-linear Fokker-Planck equations and McKean Vlasov SDEs. Then we shall recall our results on existence of weak solutions to both such equations in the singular case, where the measure dependence of the coefficients are of Nemytskii-type. The main new results to be presented are about weak uniqueness of solutions to both nonlinear Fokker-Planck equations and the corresponding McKean-Vlasov SDEs in the case of (possibly) degenerate diffusion coefficients . As a consequence of this and one obtains that the laws on path space of the solutions to the McKean-Vlasov SDEs form a nonlinear Markov process in the sense of McKean.