Universality for random matrices and log-gases

TL;DR

This paper reviews recent breakthroughs in proving the universality of local eigenvalue statistics in large random matrices and log-gases, highlighting the mechanisms and mathematical techniques involved.

Contribution

It presents the first comprehensive review of solutions to the universality conjectures for both invariant and non-invariant ensembles, emphasizing Dyson Brownian motion and regularity analysis.

Findings

Universality of local eigenvalue statistics proven for invariant ensembles.

Dyson Brownian motion's local ergodicity explains universality.

Gap distribution regularity analysis established for discrete parabolic equations.

Abstract

Eugene Wigner's revolutionary vision predicted that the energy levels of large complex quantum systems exhibit a universal behavior: the statistics of energy gaps depend only on the basic symmetry type of the model. Simplified models of Wigner's thesis have recently become mathematically accessible. For mean field models represented by large random matrices with independent entries, the celebrated Wigner-Dyson-Gaudin-Mehta (WDGM) conjecture asserts that the local eigenvalue statistics are universal. For invariant matrix models, the eigenvalue distributions are given by a log-gas with potential $V$ and inverse temperature $\beta = 1, 2, 4$. corresponding to the orthogonal, unitary and symplectic ensembles. For $\beta \not \in \{1, 2, 4\}$, there is no natural random matrix ensemble behind this model, but the analogue of the WDGM conjecture asserts that the local statistics are independent of $V$. In these lecture notes we review the recent solution to these conjectures for both invariant and non-invariant ensembles. We will discuss two different notions of universality in the sense of (i) local correlation functions and (ii) gap distributions. We will demonstrate that the local ergodicity of the Dyson Brownian motion is the intrinsic mechanism behind the universality. In particular, we review the solution of Dyson's conjecture on the local relaxation time of the Dyson Brownian motion. Additionally, the gap distribution requires a De Giorgi-Nash-Moser type H\"older regularity analysis for a discrete parabolic equation with random coefficients. Related questions such as the local version of Wigner's semicircle law and delocalization of eigenvectors will also be discussed. We will also explain how these results can be extended beyond the mean field models, especially to random band matrices.