Variance reduction using nonreversible Langevin samplers

DOI10.1007/S10955-016-1491-2zbMATH Open1343.82036arXiv1506.04934OpenAlexW2253619190WikidataQ42426717 ScholiaQ42426717MaRDI QIDQ300594FDOQ300594

G. A. Pavliotis, Andrew B. Duncan, Tony Lelièvre

Publication date: 28 June 2016

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

Abstract: A standard approach to computing expectations with respect to a given target measure is to introduce an overdamped Langevin equation which is reversible with respect to the target distribution, and to approximate the expectation by a time-averaging estimator. As has been noted in recent papers, introducing an appropriately chosen nonreversible component to the dynamics is beneficial, both in terms of reducing the asymptotic variance and of speeding up convergence to the target distribution. In this paper we present a detailed study of the dependence of the asymptotic variance on the deviation from reversibility. Our theoretical findings are supported by numerical simulations.

Full work available at URL: https://arxiv.org/abs/1506.04934

zbMATH Keywords

Langevin equation

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Central limit and other weak theorems (60F05) Stochastic methods (Fokker-Planck, Langevin, etc.) applied to problems in time-dependent statistical mechanics (82C31)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Monte Carlo strategies in scientific computing.
• Bayesian data analysis.
• Exponential convergence of Langevin distributions and their discrete approximations
• Analysis of a nonreversible Markov chain sampler.
• Langevin diffusions and Metropolis-Hastings algorithms
• Stochastic simulation: Algorithms and analysis
• Central limit theorem for additive functionals of reversible Markov processes and applications to simple exclusions
• Adaptive proposal distribution for random walk Metropolis algorithm
• Riemann Manifold Langevin and Hamiltonian Monte Carlo Methods
• Irreversible Monte Carlo algorithms for efficient sampling
• Diffusion and mixing in fluid flow
• Expansion of the global error for numerical schemes solving stochastic differential equations
• Ergodic theorems. With a supplement by Antoine Brunel
• Metastability in reversible diffusion processes. I: Sharp asymptotics for capacities and exit times
• Metastability in reversible diffusion processes. II: Precise asymptotics for small eigenvalues
• A Liapounov bound for solutions of the Poisson equation
• Stability of Markovian processes III: Foster–Lyapunov criteria for continuous-time processes
• Exponential and uniform ergodicity of Markov processes
• Simulating Hamiltonian Dynamics
• On the functional central limit theorem and the law of the iterated logarithm for Markov processes
• Spectral Theory and its Applications
• Accelerating diffusions
• Convergence of Numerical Time-Averaging and Stationary Measures via Poisson Equations
• A stochastic Newton MCMC method for large-scale statistical inverse problems with application to seismic inversion
• Subgeometric rates of convergence of \(f\)-ergodic strong Markov processes
• A central limit theorem for diffusions with periodic coefficients
• Accelerating Gaussian diffusions
• Ordering and improving the performance of Monte Carlo Markov chains.
• Multiscale diffusion processes with periodic coefficients and an application to solute transport in porous media
• On the Poisson equation and diffusion approximation. I
• Attaining the optimal Gaussian diffusion acceleration
• Variance reduction for irreversible Langevin samplers and diffusion on graphs
• Ergodicity for SDEs and approximations: locally Lipschitz vector fields and degenerate noise.
• Two Mathematical Tools to Analyze Metastable Stochastic Processes
• Central limit theorems for additive functionals of ergodic Markov diffusions processes
• Irreversible Langevin samplers and variance reduction: a large deviations approach
• Asymptotic analysis of the Green-Kubo formula
• The Effect of Mean Flows on Enhanced Diffusivity in Transport by Incompressible Periodic Velocity Fields
• 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
• Asymptotics of Solute Dispersion in Periodic Porous Media
• Control Variates for Estimation Based on Reversible Markov Chain Monte Carlo Samplers
• Variance reduction for diffusions
• Stochastic Processes and Applications
• Fluctuations in Markov Processes
• On the Construction and Comparison of Difference Schemes
• An integral representation and bounds on the effective diffusivity in passive advection by laminar and turbulent flows

Cited In (55)

• Randomized maximum likelihood based posterior sampling
• Speed up Zig-Zag
• Finite-data error bounds for Koopman-based prediction and control
• Exotic aromatic B-series for the study of long time integrators for a class of ergodic SDEs
• Generalizing Parallel Replica Dynamics: Trajectory Fragments, Asynchronous Computing, and PDMPs
• Stochastic modelling of urban structure
• Variational formulas for asymptotic variance of general discrete-time Markov chains
• Effective dynamics for non-reversible stochastic differential equations: a quantitative study
• On the Generalized Langevin Equation for Simulated Annealing
• Measuring sample quality with diffusions
• Reversible and non-reversible Markov chain Monte Carlo algorithms for reservoir simulation problems
• Non-reversible processes: GENERIC, hypocoercivity and fluctuations
• Mean field limits for interacting diffusions in a two-scale potential
• Large deviations of empirical measures of diffusions in weighted topologies
• Using perturbed underdamped Langevin dynamics to efficiently sample from probability distributions
• Modified Hamiltonian Monte Carlo for Bayesian inference
• Coarse Graining of Nonreversible Stochastic Differential Equations: Quantitative Results and Connections to Averaging
• Pressure statistics from the path integral for Darcy flow through random porous media
• NonReversible Sampling Schemes on Submanifolds
• Non-reversible metastable diffusions with Gibbs invariant measure. I: Eyring-Kramers formula
• Solving eigenvalue PDEs of metastable diffusion processes using artificial neural networks
• Some Remarks on Preconditioning Molecular Dynamics
• Long-time behaviour of degenerate diffusions: UFG-type SDEs and time-inhomogeneous hypoelliptic processes
• Response theory and phase transitions for the thermodynamic limit of interacting identical systems
• Improving the convergence of reversible samplers
• Partial differential equations and stochastic methods in molecular dynamics
• 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
• Order conditions for sampling the invariant measure of ergodic stochastic differential equations on manifolds
• Non-reversible metastable diffusions with Gibbs invariant measure. II: Markov chain convergence
• Analysis of Multiscale Integrators for Multiple Attractors and Irreversible Langevin Samplers
• Some remark on the asymptotic variance in a drift accelerated diffusion
• A Benchmark for the Bayesian Inversion of Coefficients in Partial Differential Equations
• Controlling roughening processes in the stochastic Kuramoto-Sivashinsky equation
• Variational principles for asymptotic variance of general Markov processes
• Pressure and flow statistics of Darcy flow from simulated annealing
• Ergodicity of the zigzag process
• Markov chain Monte Carlo and irreversibility
• Optimal scaling of the MALA algorithm with irreversible proposals for Gaussian targets
• Ensemble preconditioning for Markov chain Monte Carlo simulation
• Asymptotic variance for random walk Metropolis chains in high dimensions: logarithmic growth via the Poisson equation
• Accelerated convergence to equilibrium and reduced asymptotic variance for Langevin dynamics using Stratonovich perturbations
• A note on the asymptotic variance of drift accelerated diffusions
• Simulating Coulomb and log-gases with hybrid Monte Carlo algorithms
• Constructing Sampling Schemes via Coupling: Markov Semigroups and Optimal Transport
• The entropy production of stationary diffusions
• Limit theorems for the zig-zag process
• Variance reduction for Metropolis-Hastings samplers
• Interacting Langevin Diffusions: Gradient Structure and Ensemble Kalman Sampler
• Optimal friction matrix for underdamped Langevin sampling
• Structure preserving schemes for Fokker-Planck equations of irreversible processes
• Speeding up the zig-zag process
• Adaptive stepsize algorithms for Langevin dynamics
• Strong invariance principles for ergodic Markov processes

Uses Software

• BayesDA

This page was built for publication: Variance reduction using nonreversible Langevin samplers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q300594)