Freedman's inequality for matrix martingales
From MaRDI portal
(Redirected from Freedman's inequality for matrix martingales)
Abstract: Freedman's inequality is a martingale counterpart to Bernstein's inequality. This result shows that the large-deviation behavior of a martingale is controlled by the predictable quadratic variation and a uniform upper bound for the martingale difference sequence. Oliveira has recently established a natural extension of Freedman's inequality that provides tail bounds for the maximum singular value of a matrix-valued martingale. This note describes a different proof of the matrix Freedman inequality that depends on a deep theorem of Lieb from matrix analysis. This argument delivers sharp constants in the matrix Freedman inequality, and it also yields tail bounds for other types of matrix martingales. The new techniques are adapted from recent work by the present author.
Recommendations
Cited in
(42)- Time-uniform, nonparametric, nonasymptotic confidence sequences
- From Poincaré inequalities to nonlinear matrix concentration
- Nonlinear matrix concentration via semigroup methods
- Matrix concentration for products
- Freedman inequality in noncommutative probability spaces
- Nyström landmark sampling and regularized Christoffel functions
- Time-uniform Chernoff bounds via nonnegative supermartingales
- Noncommutative martingale deviation and Poincaré type inequalities with applications
- On some extensions of Bernstein's inequality for self-adjoint operators
- Matrix concentration inequalities via the method of exchangeable pairs
- A limit theorem for scaled eigenvectors of random dot product graphs
- A graphical approach to the analysis of matrix completion
- The Kőnig graph process
- Universally consistent vertex classification for latent positions graphs
- Perfect clustering for stochastic blockmodel graphs via adjacency spectral embedding
- A Local Search Framework for Experimental Design
- A Spectral Approach to Network Design
- Learning Markov models via low-rank optimization
- Online row sampling
- Randomized numerical linear algebra: Foundations and algorithms
- Strong approximation of Gaussian \(\beta\) ensemble characteristic polynomials: the hyperbolic regime
- Concentration inequalities for some negatively dependent binary random variables
- Hessian averaging in stochastic Newton methods achieves superlinear convergence
- High-dimensional estimation of quadratic variation based on penalized realized variance
- Central limit theorems for global and local empirical measures of diffusions on Erdős-Rényi graphs
- Average-case speedup for product formulas
- A noncommutative Freedman inequality
- Statistics and learning theory in the era of artificial intelligence. Abstracts from the workshop held June 23--28, 2024
- Concentration inequalities for matrix martingales in continuous time
- Matrix Freedman inequality for sub-Weibull martingales
- Online Lewis weight sampling
- Robust value iteration for optimal control of discrete-time linear systems
- User-friendly tail bounds for sums of random matrices
- Minimax learning of ergodic Markov chains
- A matrix Burkholder-Davis-Gundy inequality
- Time-uniform self-normalized concentration for vector-valued processes
- Second-order information promotes mini-batch robustness in variance-reduced gradients
- -policy gradient for online pricing
- Robust estimation of a Markov chain transition matrix from multiple sample paths
- Stochastic gradient descent in high dimensions for multi-spiked tensor PCA
- Sharp concentration for sums of matrices with Markovian dependence through universality
- Matrix Poincaré inequalities and concentration
This page was built for publication: Freedman's inequality for matrix martingales
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q638273)
