The Expected Norm of a Sum of Independent Random Matrices: An Elementary Approach

@article{Tropp2016TheEN,
  title={The Expected Norm of a Sum of Independent Random Matrices: An Elementary Approach},
  author={Joel A. Tropp},
  journal={arXiv: Probability},
  year={2016},
  pages={173-202}
}
  • J. Tropp
  • Published 15 June 2015
  • Mathematics
  • arXiv: Probability
In contemporary applied and computational mathematics, a frequent challenge is to bound the expectation of the spectral norm of a sum of independent random matrices. This quantity is controlled by the norm of the expected square of the random matrix and the expectation of the maximum squared norm achieved by one of the summands; there is also a weak dependence on the dimension of the random matrix. The purpose of this paper is to give a complete, elementary proof of this important inequality. 
The spectral norm of Gaussian matrices with correlated entries
We give a non-asymptotic bound on the spectral norm of a d×d matrix X with centered jointly Gaussian entries in terms of the covariance matrix of the entries. In some cases, this estimate is sharp
Matrix Concentration Inequalities and Free Probability
TLDR
These bounds quantify the degree to which the deterministic matrices Ai behave as though they are freely independent, and imply strong asymptotic freeness for a remarkably general class of Gaussian random matrix models, including matrices that may be very sparse and that lack any special symmetries.
PR ] 1 3 A ug 2 02 1 MATRIX CONCENTRATION INEQUALITIES AND FREE PROBABILITY
TLDR
These bounds quantify the degree to which the deterministic matrices Ai behave as though they are freely independent, and imply strong asymptotic freeness for a remarkably general class of Gaussian random matrix models, including matrices that may be very sparse and that lack any special symmetries.
Moment inequalities for matrix-valued U-statistics of order 2
We present Rosenthal-type moment inequalities for matrix-valued U-statistics of order 2. As a corollary, we obtain new matrix concentration inequalities for U-statistics. One of our main technical
From Poincaré inequalities to nonlinear matrix concentration
This paper deduces exponential matrix concentration from a Poincare inequality via a short, conceptual argument. Among other examples, this theory applies to matrix-valued functions of a uniformly
Operator Norm Moment and Exponential Inequalities for Matrix U-statistics
We present a Rosenthal-type moment bound and a Bernstein-type exponential tail bound for order 1 degenerated matrix U-statistics in operator norm, with explicit factors on dimension dependencies. We
Matrix Concentration for Products
This paper develops nonasymptotic growth and concentration bounds for a product of independent random matrices. These results sharpen and generalize recent work of Henriksen-Ward, and they are
High-Dimensional Probability
TLDR
A broad range of illustrations is embedded throughout, including classical and modern results for covariance estimation, clustering, networks, semidefinite programming, coding, dimension reduction, matrix completion, machine learning, compressed sensing, and sparse regression.
Randomized numerical linear algebra: Foundations and algorithms
TLDR
This survey describes probabilistic algorithms for linear algebraic computations, such as factorizing matrices and solving linear systems, that have a proven track record for real-world problems and treats both the theoretical foundations of the subject and practical computational issues.
Universality laws for randomized dimension reduction, with applications
TLDR
It is proved that there is a phase transition in the success probability of the dimension reduction map as the embedding dimension increases, and each map has the same stability properties, as quantified through the restricted minimum singular value.
...
...

References

SHOWING 1-10 OF 73 REFERENCES
An Introduction to Matrix Concentration Inequalities
  • J. Tropp
  • Mathematics
    Found. Trends Mach. Learn.
  • 2015
TLDR
The aim of this monograph is to describe the most successful methods from this area along with some interesting examples that these techniques can illuminate.
Matrix concentration inequalities via the method of exchangeable pairs
This paper derives exponential concentration inequalities and polynomial moment inequalities for the spectral norm of a random matrix. The analysis requires a matrix extension of the scalar
The lower tail of random quadratic forms with applications to ordinary least squares
TLDR
This paper proves that the “lower tail” of such a matrix is sub-Gaussian under a simple fourth moment assumption on the one-dimensional marginals of the random vectors, and obtains a nearly optimal finite-sample result for the ordinary least squares estimator under random design.
Topics in Random Matrix Theory
The field of random matrix theory has seen an explosion of activity in recent years, with connections to many areas of mathematics and physics. However, this makes the current state of the field
Concentration Inequalities - A Nonasymptotic Theory of Independence
TLDR
Deep connections with isoperimetric problems are revealed whilst special attention is paid to applications to the supremum of empirical processes.
On the subspaces ofLp(p>2) spanned by sequences of independent random variables
Let 2<p<∞. The Banach space spanned by a sequence of independent random variables inLp, each of mean zero, is shown to be isomorphic tol2,lp,l2⊕lp, or a new spaceXp, and the linear topological
Noncommutative Burkholder/Rosenthal inequalities II: Applications
We show norm estimates for the sum of independent random variables in noncommutative Lp-spaces for 1 < p < ∞, following our previous work. These estimates generalize the classical Rosenthal
Random Vectors in the Isotropic Position
Abstract Letybe a random vector in R n, satisfying E y⊗y=id. LetMbe a natural number and lety1, …, yMbe independent copies ofy. We study the question of approximation of the identity operator by
The spectrum of random k-lifts of large graphs (with possibly large k)
We study random k-lifts of large, but otherwise arbitrary graphs G. We prove that, with high probability, all eigenvalues of the adjacency matrix of the lift that are not eigenvalues of G are of the
...
...