Steep points of Gaussian free fields in any dimension (Q2664529)

The author investigates the Hausdorff dimension of a general class of exceptional sets of a Gaussian free field (GFF). Log-correlated and polynomial-correlated GFFs in any dimension are studied. The approach is based on a circle or sphere averaging regularization. The main tool is the notion of ``\(f\)-steep points'' of the GFF, i.e., locations where the regularized field changes very fast if the weight function is \(f.\) The paper investigates the Hausdorff dimension of the set of \(f\)-steep points. Lower and upper bounds on the Hausdorff dimension of the exceptional sets are derived. Existing results on thick point sets for both log-correlated and polynomial-correlated GFFs are obtained as a corollary. Also, results for new exceptional sets are given. Several examples how various choices of \(f\) lead to different exceptional sets of GFFs are provided. Some related research problems and future directions are presented in the last section.