Repeated singular values of a random symmetric matrix and decoupled singular value estimates

Abstract: Let $A_n$ be a random symmetric matrix with Bernoulli ${\pm 1}$ entries. For any $\kappa>0$ and two real numbers $\lambda_1,\lambda_2$ with a separation $|\lambda_1-\lambda_2|\geq \kappa n^{1/2}$ and both lying in the bulk $[-(2-\kappa)n^{1/2},(2-\kappa)n^{1/2}]$, we prove a joint singular value estimate $$ \mathbb{P}(\sigma_{min}(A_n-\lambda_i I_n)\leq\epsilon n^{-1/2};i=1,2)\leq C\epsilon^2+2e^{-cn}. $$ For general subgaussian distribution and a mesoscopic separation $|\lambda_1-\lambda_2|\geq \kappa n^{-1/2+\sigma},\sigma>0$ we prove the same estimate with $e^{-cn}$ replaced by an exponential type error. This means that extreme behaviors of the least singular value at two locations can essentially be decoupled all the way down to the exponential scale when the two locations are separated. As a corollary, we prove that all the singular values of $A_n$ in $[\kappa n^{1/2},(2-\kappa)n^{1/2}]$ are distinct with probability $1-e^{-cn}$, and with high probability the minimal gap between these singular values has order at least $n^{-3/2}$. This justifies, in a strong quantitative form, a conjecture of Vu up to $(1-\kappa)$-fraction of the spectrum for any $\kappa>0$.