English

Norms of structured random matrices

Probability 2024-11-19 v2 Functional Analysis

Abstract

For m,nNm,n\in\mathbb{N} let X=(Xij)im,jnX=(X_{ij})_{i\leq m,j\leq n} be a random matrix, A=(aij)im,jnA=(a_{ij})_{i\leq m,j\leq n} a real deterministic matrix, and XA=(aijXij)im,jnX_A=(a_{ij}X_{ij})_{i\leq m,j\leq n} the corresponding structured random matrix. We study the expected operator norm of XAX_A considered as a random operator between pn\ell_p^n and qm\ell_q^m for 1p,q1\leq p,q \leq \infty. We prove optimal bounds up to logarithmic terms when the underlying random matrix XX has i.i.d. Gaussian entries, independent mean-zero bounded entries, or independent mean-zero ψr\psi_r (r(0,2]r\in(0,2]) entries. In certain cases, we determine the precise order of the expected norm up to constants. Our results are expressed through a sum of operator norms of Hadamard products AAA\circ A and (AA)T(A\circ A)^T.

Cite

@article{arxiv.2112.14413,
  title  = {Norms of structured random matrices},
  author = {Radosław Adamczak and Joscha Prochno and Marta Strzelecka and Michał Strzelecki},
  journal= {arXiv preprint arXiv:2112.14413},
  year   = {2024}
}

Comments

50 pages, 1 figure, 1 table; Remark 1.1 and Subsection 5.4 added, typos corrected

R2 v1 2026-06-24T08:34:21.613Z