BBP Transition for Deformed Inhomogeneous Random Matrices

Source abstract

We study finite-rank deformations of inhomogeneous random symmetric/Hermitian matrices $H_N=(σ_{ij}W_{ij})_{i,j=1}^N$, whose variance profile satisfies $\sum_jσ_{ij}^2=1$. Writing $σ_N^*=\max_{i,j}σ_{ij}$ for the matrix sparsity parameter, we establish the BBP transition of outlier eigenvalues at the level of the law of large numbers in two regimes of moment conditions on $\{W_{ij}\}$: (1) under a uniform sub-Weibull bound with exponent $0<α\le 2$ and the sharp assumption $σ_N^*(\log N)^{1/α}\to 0$, the convergence holds almost surely; (2) under uniformly integrable $p$-th powers for some $p\ge 4$ and the sharp assumption $N(σ_N^*)^{p-2}=O(1)$, the convergence holds in probability. Our approach combines sharp norm estimates, fixed-degree polynomial isotropic laws, and a finite-dimensional resolvent equation.