Id: arxiv/2602.17960 Url: https://arxiv.org/abs/2602.17960 Tags: math.PR, stat.AP, stat.ML Score: 0 Title: Anisotropic local law for non-separable sample covariance matrices Domain: math.PR Source: arxiv Authors: Zhou Fan, Renyuan Ma, Elliot Paquette, Zhichao Wang Reviews: 0 Subject: math.PR Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-02-20 Page: https://delphi.science/r/arxiv/2602.17960 We establish local laws for sample covariance matrices $K = N^{-1}\sum_{i=1}^N \g_i\g_i^*$ where the random vectors $\g_1, \ldots, \g_N \in \R^n$ are independent with common covariance $Σ$. Previous work has largely focused on the separable model $\g = Σ^{1/2}\w$ with $\w$ having independent entries, but this structure is rarely present in statistical applications involving dependent or nonlinearly transformed data. Under a concentration assumption for quadratic forms $\g^*A\g$, we prove an optimal averaged local law showing that the Stieltjes transform of $K$ converges to its deterministic…