A unified approach for large deviations of bulk and extreme eigenvalues of the Wishart ensemble
Adolfo Camacho Melo, Isaac P\'erez Castillo

TL;DR
This paper introduces a unified Coulomb fluid-based method to analyze large deviations of both bulk and extreme eigenvalues in Wishart matrices, providing a comprehensive rate function and validating results with simulations.
Contribution
It develops a unified analytical framework to derive large deviation rate functions for bulk and extreme eigenvalues of Wishart matrices, connecting shifted index statistics with eigenvalue deviations.
Findings
Derived a rate function $ ext{Psi}(c, x)$ for large deviations
Connected bulk and edge eigenvalue deviations through limits
Validated analytical results with Monte Carlo simulations
Abstract
Within the framework of the Coulomb fluid picture, we present a unified approach to derive the large deviations of bulk and extreme eigenvalues of large Wishart matrices. By analysing the statistics of the shifted index number we are able to derive a rate function depending on two variables: the fraction of eigenvalues to the left of an infinite energetic barrier at position . For a fixed value of , the rate function gives the large deviations of the bulk eigenvalues. In particular, in the limits or it is possible to extract the left and right deviations of the smallest and largest eigenvalues, respectively. Alternatively, for a fixed value of the barrier, the rate function provides the large deviations of the shifted index number. All our analytical findings are compared with Metropolis Monte Carlo simulations, obtaining excellent agreement.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRandom Matrices and Applications · Theoretical and Computational Physics · Statistical Methods and Bayesian Inference
