Universality and sharp matrix concentration inequalities

Authors Tatiana Brailovskaya, Ramon van Handel

Publication date 13 January 2022

Random matrices (algebraic aspects) (15B52) Noncommutative probability and statistics (46L53) Free probability and free operator algebras (46L54) Random matrices (probabilistic aspects) (60B20) Inequalities; stochastic orderings (60E15)

Abstract: We show that, under mild assumptions, the spectrum of a sum of independent random matrices is close to that of the Gaussian random matrix whose entries have the same mean and covariance. This nonasymptotic universality principle both explains the phenomenon behind classical matrix concentration inequalities such as the matrix Bernstein inequality, and yields new sharp matrix concentration inequalities for general sums of independent random matrices when combined with the recent Gaussian theory of Bandeira, Boedihardjo, and Van Handel. As an application of our main results, we prove strong asymptotic freeness of a very general class of random matrix models with non-Gaussian, nonhomogeneous, and dependent entries.

This page was built for publication: Universality and sharp matrix concentration inequalities

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6388209)