Skip to Main content Skip to Navigation
Theses

Statistical mechanics of Gaussian fields

Abstract : In this thesis, we study the level sets of smooth Gaussian fields, or random smooth functions. Several directions are explored, some linked to spectral theory, some to statistical mechanics.The first object of focus is a family of Gaussian fields on compact Riemannian manifolds defined as linear combinations of eigenfunctions of the Laplacian with independent Gaussian weights. In special cases, this family specializes to the band-limited ensemble which has received a lot of attention in recent years, but also to the cut-off Gaussian Free Field, which is the projection of the Gaussian Free Field on the first eigenspaces of the Laplacian. We study the covariance function of these fields, the expected number of connected components of their zero set, and, in the case of the cut-off Gaussian Free Field, derive a precise large deviation estimate on the event that the field is positive on a fixed set when the energy cut-off tends to infinity.Next, we study percolation of excursion sets of stationary fields on the plane using techniques from Bernoulli precolation. We first derive a mixing bound for the topology of nodal sets of planar Gaussian fields. Then, we prove a sharp phase transition result for the Bargmann-Fock random field.
Complete list of metadatas

Cited literature [183 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02078812
Contributor : Abes Star :  Contact
Submitted on : Monday, March 25, 2019 - 4:05:09 PM
Last modification on : Friday, November 6, 2020 - 4:26:57 AM
Long-term archiving on: : Wednesday, June 26, 2019 - 3:21:26 PM

File

RIVERA_2018_archivage.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02078812, version 1

Collections

Citation

Alejandro Rivera. Statistical mechanics of Gaussian fields. Statistical Mechanics [cond-mat.stat-mech]. Université Grenoble Alpes, 2018. English. ⟨NNT : 2018GREAM066⟩. ⟨tel-02078812⟩

Share