May 2024 Large deviations for random matrices in the orthogonal group and Stiefel manifold with applications to random projections of product distributions
Zakhar Kabluchko, Joscha Prochno
Author Affiliations +
Ann. Inst. H. Poincaré Probab. Statist. 60(2): 990-1024 (May 2024). DOI: 10.1214/22-AIHP1340

Abstract

We prove large deviation principles (LDPs) for random matrices in the orthogonal group and Stiefel manifold, determining both the speed and good convex rate functions that are explicitly given in terms of certain log-determinants of trace-class operators and are finite on the set of Hilbert-Schmidt operators M satisfying MM<1. As an application of those LDPs, we determine the precise large deviation behavior of k-dimensional random projections of high-dimensional product distributions using an appropriate interpretation in terms of point processes, also characterizing the space of all possible deviations. The case of uniform distributions on p-balls, 1p, is then considered and reduced to appropriate product measures. Those applications generalize considerably the recent work (Studia Mathematica 264 (2022) 103–119).

Nous prouvons des principes de grandes déviations (LDPs) pour les matrices aléatoires uniformes sur le groupe orthogonal et les variétés de Stiefel, en déterminant à la fois la vitesse et les bonnes fonctions de taux convexes qui sont explicitement données en termes de certains log-déterminants d’opérateurs à trace, et sont finies sur l’ensemble des opérateurs de Hilbert-Schmidt M satisfaisant MM<1. Comme application de ces LDPs, nous déterminons le comportement précis des grandes déviations des projections aléatoires de dimension k des lois de produit de grande dimension en utilisant une interprétation appropriée en termes de processus ponctuels, caractérisant également l’espace de toutes les déviations possibles. Le cas des lois uniformes sur les boules p, 1p, est ensuite considéré et réduit à des mesures produit appropriées. Ces applications généralisent considérablement les travaux récents (Studia Mathematica 264 (2022) 103–119).

Funding Statement

Zakhar Kabluchko has been supported by the German Research Foundation under Germany’s Excellence Strategy EXC 2044 – 390685587, Mathematics Münster: Dynamics – Geometry – Structure and by the DFG priority program SPP 2265 Random Geometric Systems.
Joscha Prochno is supported by the Austrian Science Fund (FWF) Project P32405 Asymptotic Geometric Analysis and Applications and by the FWF Project F5513-N26 which is a part of the Special Research Program Quasi-Monte Carlo Methods: Theory and Applications.

Acknowledgements

We are grateful to Gerold Alsmeyer for drawing our attention to the works of J. V. Linnik related to Lemma 6.4.

Citation

Download Citation

Zakhar Kabluchko. Joscha Prochno. "Large deviations for random matrices in the orthogonal group and Stiefel manifold with applications to random projections of product distributions." Ann. Inst. H. Poincaré Probab. Statist. 60 (2) 990 - 1024, May 2024. https://doi.org/10.1214/22-AIHP1340

Information

Received: 1 March 2022; Revised: 16 September 2022; Accepted: 28 October 2022; Published: May 2024
First available in Project Euclid: 11 June 2024

Digital Object Identifier: 10.1214/22-AIHP1340

Subjects:
Primary: 52A23 , 60B20 , 60F10
Secondary: 46B06 , 52A22

Keywords: large deviation principle , matrix variate distributions , orthogonal group , Product measure , projective limit , Random matrix , random projection , Stiefel manifold

Rights: Copyright © 2024 Association des Publications de l’Institut Henri Poincaré

Vol.60 • No. 2 • May 2024
Back to Top