Dimension-free tail inequalities for sums of random matrices

Abstract

We derive exponential tail inequalities for sums of random matrices with no dependence on the explicit matrix dimensions. These are similar to the matrix versions of the Chernoff bound and Bernstein inequality except with the explicit matrix dimensions replaced by a trace quantity that can be small even when the dimension is large or infinite. Some applications to principal component analysis and approximate matrix multiplication are given to illustrate the utility of the new bounds.

References in corpus (5)

• Quantum state tomography via compressed sensing
• Consistency of the group Lasso and multiple kernel learning
• Concentration of the adjacency matrix and of the Laplacian in random graphs with independent edges
• Large deviations and stochastic calculus for large random matrices
• Freedman's inequality for matrix martingales

Cited by in corpus (5)

• Spectral Methods for Learning Multivariate Latent Tree Structure
• UPAL: Unbiased Pool Based Active Learning
• Excess risk bounds for multitask learning with trace norm regularization
• Dimension-free Concentration Bounds on Hankel Matrices for Spectral Learning
• On Dimension-free Tail Inequalities for Sums of Random Matrices and Applications