Strong approximation of Gaussian \(\beta\) ensemble characteristic polynomials: the hyperbolic regime (Q6103977): Difference between revisions

In this article, the authors study the characteristic polynomials \(\varphi_{N}\) of the Gaussian \(\beta\)-ensemble for general \(\beta>0\), and extend its known connections to log-correlated fields. In order to describe the asymptotics of the characteristic polynomial, the authors distinguish the three regimes (the hyperbolic, parabolic and elliptic) of transfer matrix recurrences of Gaussian \(\beta\)-ensemble and analyze completely the hyperbolic part of the recurrence. Applying the strong embedding theorem, the authors establish a new coupling between \(\varphi_{N}\) and a Gaussian analytic function with an error which is uniform away from the support of the semicircle law. Borrowing from the moderate deviation estimates and transfer matrix recurrence, the authors further obtain a (probabilistic) approximation for \(\varphi_{N}\) in terms of a Gaussian log-correlated field in the hyperbolic region, that is, the Gaussian \(\beta\)-ensemble characteristic polynomials can be approximated by the Hermite polynomial times the exponential of a Gaussian analytic function which arises by linearizing the transfer matrix recurrence in the hyperbolic region.