Subcritical multiplicative chaos for regularized counting statistics from random matrix theory (Q1637282)

Consider a random unitary \(N\times N\)-matrix \(U_N\) sampled from the unitary group according to the Haar measure (circular unitary ensemble). The authors study a random process \(X_N\) counting essentially the number of eigenvalues of \(U_N\) in a mesoscopic arc centered at the point \(u\) on the unit circle, after certain smoothing. They prove that the exponential of this process, namely \(e^{\gamma X_N}\), converges weakly to the multiplicative Gaussian chaos, after suitable normalization. The proof is valid in the whole subcritical phase \(0 < \gamma < \sqrt 2\) of the Gaussian multiplicative chaos. The authors discuss implications of this result for obtaining a lower bound on the maximum of \(X_N\). They also show that the moments of the total mass of \(e^{\gamma X_n}\) converge to Selberg-like integrals and establish analogous results with eigenvalues replaced by the points of the sine process.