Universality conjecture for diagonal resolvent entries of rotationally invariant random matrices

Let \vbM\vb{M} be a rotationally invariant random matrix of size NN with a non-pathological limiting spectral density. For zz outside the spectrum of \vbM\vb{M}, define

[\vbG\vbMN(z)]11:=[(z\vb1\vbM)1]11.[\vb{G}_{\vb{M}}^N(z)]_{11}:=[(z\vb{1}-\vb{M})^{-1}]_{11}.

Let g\vbM(z)\mathfrak{g}_{\vb{M}}(z) denote the limiting Stieltjes transform and let O~(z)\widetilde{\mathcal{O}}(z) denote the rescaled self-overlap corresponding to zz. Universality conjecture. As NN\to\infty, the distribution of [\vbG\vbMN(z)]11[\vb{G}_{\vb{M}}^N(z)]_{11} converges in law to a complex Student law with ν=2\nu=2 and density

ω1πβ(β+ωg\vbM(z))2,β=πO~(z).\omega\mapsto\frac{1}{\pi}\frac{\beta}{(\beta+|\omega-\mathfrak{g}_{\vb{M}}(z)|)^2},\qquad \beta=\pi\widetilde{\mathcal{O}}(z).

Equivalently, the distribution of

Ω:=[\vbG\vbMN(z)]11g\vbM(z)πO~(z)\Omega:=\frac{[\vb{G}_{\vb{M}}^N(z)]_{11}-\mathfrak{g}_{\vb{M}}(z)}{\sqrt{\pi\widetilde{O}(z)}}

should be independent of \vbM\vb{M} and zz, with density ω1π(1+ω2)2\omega\mapsto \frac{1}{\pi(1+|\omega|^2)^2}. This conjecture proposes a universal limiting law for diagonal resolvent entries beyond the Ginibre case; its resolution is not established by the supplied text.

Sources & referencesView supporting material

Primary source

Pierre Bousseyroux, Jean-Philippe Bouchaud and Marc Potters, “Distribution of the Diagonal Entries of the Resolvent of a Complex Ginibre Matrix”, arXiv:2411.19266 (2024).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.