Title:Large Deviations of Extreme Eigenvalues of Generalized Sample Covariance Matrices

Abstract:Very rare events in which the largest eigenvalue of a random matrix is atypically large have important consequences in statistics, e.g. in principal components analysis, and for studying the rough high-dimensional landscapes encountered in disordered systems in statistical mechanics. These problems lead to consider matrices $(1/m) \sum_{\mu=1}^m d_\mu \mathbf{z}_\mu \mathbf{z}_\mu^\dagger$, with $\{\mathbf{z}_\mu\}_{\mu=1}^m$ standard Gaussian vectors of size $n$, and (fixed) real $d_\mu$. In a high-dimensional limit we leverage recent techniques to derive the probability of large deviations of the extreme eigenvalues away from the bulk. We probe our results with Monte-Carlo methods that effectively simulate events with probability as small as $10^{-100}$.