Estimating covariances in spin systems
The task of computing the covariance matrix of an Ising model has recently emerged as an important component in establishing functional inequalities and rapid mixing of associated Markov chains. In this talk we will explain a general method to represent these matrices. In certain cases, for example when the underlying graph is an expander, we can combine this representation with normal approximation techniques to produce extremely tight estimates.
Joint work with Youngtak Sohn