English

Sparse matrices: convergence of the characteristic polynomial seen from infinity

Probability 2021-06-23 v2

Abstract

We prove that the reverse characteristic polynomial det(InzAn)\det(I_n - zA_n) of a random n×nn \times n matrix AnA_n with iid Bernoulli(d/n)\mathrm{Bernoulli}(d/n) entries converges in distribution towards the random infinite product =1(1z)Y\prod_{\ell = 1}^\infty(1-z^\ell)^{Y_\ell} where YY_\ell are independent Poisson(d/)\mathrm{Poisson}(d^\ell/\ell) random variables. We show that this random function is a Poisson analog of more classical Gaussian objects such as the Gaussian holomorphic chaos. As a byproduct, we obtain new simple proofs of previous results on the asymptotic behaviour of extremal eigenvalues of sparse Erd\H{o}s-R\'enyi digraphs: for every d>1d>1, the greatest eigenvalue of AnA_n is close to dd and the second greatest is smaller than d\sqrt{d}, a Ramanujan-like property for irregular digraphs. For d<1d<1, the only non-zero eigenvalues of AnA_n converge to a Poisson multipoint process on the unit circle. Our results also extend to the semi-sparse regime where dd is allowed to grow to \infty with nn, slower than no(1)n^{o(1)}. We show that the reverse characteristic polynomial converges towards a more classical object written in terms of the exponential of a log-correlated real Gaussian field, as in the dense case studied in a recent paper \cite{bordenave2020convergence}. In the semi-sparse regime, the empirical spectral distribution of An/dnA_n/\sqrt{d_n} converges to the circle distribution; as a consequence of our results, the second eigenvalue sticks to the edge of the circle.

Keywords

Cite

@article{arxiv.2106.00593,
  title  = {Sparse matrices: convergence of the characteristic polynomial seen from infinity},
  author = {Simon Coste},
  journal= {arXiv preprint arXiv:2106.00593},
  year   = {2021}
}

Comments

Added the semi-sparse case $d_n \to \infty$ and a more complete description of the $d_n=d<1$ case

R2 v1 2026-06-24T02:42:57.526Z