Improving the convergence of reversible samplers

DOI10.1007/S10955-016-1565-1zbMATH Open1348.82046arXiv1601.08118OpenAlexW2257408096WikidataQ57453402 ScholiaQ57453402MaRDI QIDQ330615FDOQ330615

Authors: Luc Rey-Bellet, Konstantinos Spiliopoulos 

Publication date: 26 October 2016

Published in: Journal of Statistical Physics (Search for Journal in Brave)

Abstract: In Monte-Carlo methods the Markov processes used to sample a given target distribution usually satisfy detailed balance, i.e. they are time-reversible. However, relatively recent results have demonstrated that appropriate reversible and irreversible perturbations can accelerate convergence to equilibrium. In this paper we present some general design principles which apply to general Markov processes. Working with the generator of Markov processes, we prove that for some of the most commonly used performance criteria, i.e., spectral gap, asymptotic variance and large deviation functionals, sampling is improved for appropriate reversible and irreversible perturbations of some initially given reversible sampler. Moreover we provide specific constructions for such reversible and irreversible perturbations for various commonly used Markov processes, such as Markov chains and diffusions. In the case of diffusions, we make the discussion more specific using the large deviations rate function as a measure of performance.

Full work available at URL: https://arxiv.org/abs/1601.08118

Recommendations

• Irreversible Langevin samplers and variance reduction: a large deviations approach
• Accelerating reversible Markov chains
• Analysis of a nonreversible Markov chain sampler.
• Markov chain Monte Carlo and irreversibility
• On the convergence time of some non-reversible Markov chain Monte Carlo methods

zbMATH Keywords

Monte Carlo samplinglarge deviationsasymptotic varianceMarkov processesdetailed balanceirreversibilityLangevin sampling

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Numerical analysis or methods applied to Markov chains (65C40) Large deviations (60F10) Diffusion processes (60J60) Combinatorial probability (60C05)

Cites Work

• Markov Chains and Stochastic Stability
• Asymptotic evaluation of certain markov process expectations for large time, I
• Analysis of a nonreversible Markov chain sampler.
• A note on Metropolis-Hastings kernels for general state spaces
• Stochastic simulation: Algorithms and analysis
• General state space Markov chains and MCMC algorithms
• Rates of convergence of the Hastings and Metropolis algorithms
• Riemann Manifold Langevin and Hamiltonian Monte Carlo Methods
• Irreversible Monte Carlo algorithms for efficient sampling
• Title not available (Why is that?)
• Large deviations
• Accelerating diffusions
• Accelerating Gaussian diffusions
• Ordering and improving the performance of Monte Carlo Markov chains.
• Variance reduction for irreversible Langevin samplers and diffusion on graphs
• Irreversible Langevin samplers and variance reduction: a large deviations approach
• Variance reduction using nonreversible Langevin samplers
• Optimum Monte-Carlo sampling using Markov chains
• Non-reversible Metropolis-Hastings
• The behavior of the spectral gap under growing drift
• Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion
• Variance reduction for diffusions
• Optimal spectral structure of reversible stochastic matrices, Monte Carlo methods and the simulation of Markov random fields
• On the convergence of the Markov chain simulation method
• On the large deviation rate function for the empirical measures of reversible jump Markov processes
• On the spectral analysis of second-order Markov chains
• Accelerating reversible Markov chains
• An extension of Peskun and Tierney orderings to continuous time Markov chains
• On Large Deviations from the Invariant Measure
• Title not available (Why is that?)
• Title not available (Why is that?)
• On the infinite swapping limit for parallel tempering
• Monotonicity of the dynamical activity
• Efficiency of finite state space Monte Carlo Markov chains

Cited In (21)

• Gibbsian stationary non-equilibrium states
• Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion
• Large deviations of empirical measures of diffusions in weighted topologies
• Characterizing limits and opportunities in speeding up Markov chain mixing
• NonReversible Sampling Schemes on Submanifolds
• Non-reversible metastable diffusions with Gibbs invariant measure. I: Eyring-Kramers formula
• Large deviations for the skew-detailed-balance lifted-Markov processes to sample the equilibrium distribution of the Curie–Weiss model
• Is there an analog of Nesterov acceleration for gradient-based MCMC?
• Optimal Variance Reduction for Markov Chain Monte Carlo
• Acceleration of convergence to equilibrium in Markov chains by breaking detailed balance
• Irreversible samplers from jump and continuous Markov processes
• On Irreversible Metropolis Sampling Related to Langevin Dynamics
• Peskun-Tierney ordering for Markovian Monte Carlo: beyond the reversible scenario
• Non-reversible metastable diffusions with Gibbs invariant measure. II: Markov chain convergence
• Analysis of Multiscale Integrators for Multiple Attractors and Irreversible Langevin Samplers
• Geometry-informed irreversible perturbations for accelerated convergence of Langevin dynamics
• Accelerated convergence to equilibrium and reduced asymptotic variance for Langevin dynamics using Stratonovich perturbations
• A large deviation principle for the empirical measures of Metropolis-Hastings chains
• Irreversible Monte Carlo algorithms for efficient sampling
• Constructing Sampling Schemes via Coupling: Markov Semigroups and Optimal Transport
• Limit theorems for the zig-zag process

This page was built for publication: Improving the convergence of reversible samplers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q330615)