Approximate spectral gaps for Markov chain mixing times in high dimensions
From MaRDI portal
(Redirected from Publication:5154639)
Abstract: This paper introduces a concept of approximate spectral gap to analyze the mixing time of Markov Chain Monte Carlo (MCMC) algorithms for which the usual spectral gap is degenerate or almost degenerate. We use the idea to analyze a class of MCMC algorithms to sample from mixtures of densities. As an application we study the mixing time of a Gibbs sampler for variable selection in linear regression models. Under some regularity conditions on the signal and the design matrix of the regression problem, we show that for well-chosen initial distributions the mixing time of the Gibbs sampler is polynomial in the dimension of the space.
Recommendations
- Approximate bounding of mixing time for multiple-step Gibbs samplers
- On the computational complexity of high-dimensional Bayesian variable selection
- Function-specific mixing times and concentration away from equilibrium
- scientific article; zbMATH DE number 7249109
- Spectral gaps for a Metropolis-Hastings algorithm in infinite dimensions
Cites work
- \(L_ 2\) rates of convergence for attractive reversible nearest particle systems: The critical case
- Bayesian variable selection with shrinking and diffusing priors
- Conditions for rapid mixing of parallel and simulated tempering on multimodal distributions
- Convergence of MCMC and loopy BP in the tree uniqueness region for the hard-core model
- Elementary bounds on Poincaré and log-Sobolev constants for decomposable Markov chains
- Error bounds for Metropolis-Hastings algorithms applied to perturbations of Gaussian measures in high dimensions
- Geometric bounds for eigenvalues of Markov chains
- Hit-and-run mixes fast
- scientific article; zbMATH DE number 1034042 (Why is no real title available?)
- Improved Bounds for Mixing Rates of Markov Chains and Multicommodity Flow
- Lasso-type recovery of sparse representations for high-dimensional data
- Log-concave sampling: Metropolis-Hastings algorithms are fast
- Markov chain decomposition for convergence rate analysis
- Mathematical aspects of mixing times in Markov chains.
- Nonasymptotic mixing of the MALA algorithm
- On the computational complexity of high-dimensional Bayesian variable selection
- Perturbation bounds for Monte Carlo within metropolis via restricted approximations
- Random walks in a convex body and an improved volume algorithm
- Sampling from log-concave distributions
- The geometry of logconcave functions and sampling algorithms
- Trends to equilibrium in total variation distance
Cited in
(14)- Spectral gap of replica exchange Langevin diffusion on mixture distributions
- Approximate bounding of mixing time for multiple-step Gibbs samplers
- Function-specific mixing times and concentration away from equilibrium
- Spectral gaps for a Metropolis-Hastings algorithm in infinite dimensions
- Approximate verification of geometric ergodicity for multiple-step Metropolis transition kernels
- Markov Kernels Local Aggregation for Noise Vanishing Distribution Sampling
- An asymptotic Peskun ordering and its application to lifted samplers
- Mixing of Metropolis-adjusted Markov chains via couplings: the high acceptance regime
- Dimension-free mixing times of Gibbs samplers for Bayesian hierarchical models
- Dimension-free relaxation times of informed MCMC samplers on discrete spaces
- Spectral gap bounds for reversible hybrid Gibbs chains
- Entropy-based burn-in time analysis and ranking for (A)MCMC algorithms in high dimension
- Scalability of Metropolis-within-Gibbs schemes for high-dimensional Bayesian models
- Mixing time bounds via the spectral profile
This page was built for publication: Approximate spectral gaps for Markov chain mixing times in high dimensions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5154639)
