Variance Reduction using Nonreversible Langevin Samplers

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.

30 pages

References in corpus (13)

• On the Poisson equation and diffusion approximation 3
• A Complete Recipe for Stochastic Gradient MCMC
• Irreversible Monte Carlo Algorithms for Efficient Sampling
• Optimal non-reversible linear drift for the convergence to equilibrium of a diffusion
• Accelerating diffusions
• Convergence of Numerical Time-Averaging and Stationary Measures via Poisson Equations
• Irreversible Langevin samplers and variance reduction: a large deviation approach
• Improving Asymptotic Variance of MCMC Estimators: Non-reversible Chains are Better
• Improving the Asymptotic Performance of Markov Chain Monte-Carlo by Inserting Vortices
• Central limit theorems for additive functionals of ergodic Markov diffusions processes
• Control of probability flow in Markov chain Monte Carlo -- Nonreversibility and lifting
• Langevin dynamics neglecting detailed balance condition
• Variance reduction for diffusions

Cited by in corpus (47)

• The Zig-Zag Process and Super-Efficient Sampling for Bayesian Analysis of Big Data
• Ergodicity of the zigzag process
• A piecewise deterministic scaling limit of Lifted Metropolis-Hastings in the Curie-Weiss model
• Improving the convergence of reversible samplers
• Mean Field Limits for Interacting Diffusions in a Two-Scale Potential
• Modified Hamiltonian Monte Carlo for Bayesian inference
• Acceleration of convergence to equilibrium in Markov chains by breaking detailed balance
• Limit theorems for the Zig-Zag process
• Sampling efficiency of transverse forces in dense liquids
• Using Perturbed Underdamped Langevin Dynamics to Efficiently Sample from Probability Distributions
• Response Theory and Phase Transitions for the Thermodynamic Limit of Interacting Identical Systems
• Role of current fluctuations in nonreversible samplers
• Stochastic Modelling of Urban Structure
• Order conditions for sampling the invariant measure of ergodic stochastic differential equations on manifolds
• Solving eigenvalue PDEs of metastable diffusion processes using artificial neural networks
• Breaking Reversibility Accelerates Langevin Dynamics for Global Non-Convex Optimization
• Controlling roughening processes in the stochastic Kuramoto-Sivashinsky equation
• Exotic aromatic B-series for the study of long time integrators for a class of ergodic SDEs
• Nonreversible Langevin Samplers: Splitting Schemes, Analysis and Implementation
• Simulating Coulomb gases and log-gases with hybrid Monte Carlo algorithms
• Irreversible Samplers from Jump and Continuous Markov Processes
• Noise-induced transitions in rugged energy landscapes
• Effective dynamics for non-reversible stochastic differential equations: a quantitative study
• Dynamics of the Desai-Zwanzig model in multi-well and random energy landscapes
• Randomized maximum likelihood based posterior sampling
• Generalized and hybrid Metropolis-Hastings overdamped Langevin algorithms
• Accelerating, to some extent, the $p$-spin dynamics
• Irreversible Boltzmann samplers in dense liquids: weak-coupling approximation and mode-coupling theory
• Geometric Methods for Sampling, Optimisation, Inference and Adaptive Agents
• Optimal friction matrix for underdamped Langevin sampling
• On the Generalised Langevin Equation for Simulated Annealing
• Control of probability flow in Markov chain Monte Carlo -- Nonreversibility and lifting
• The universal equivariance properties of exotic aromatic B-series
• Concentration Inequalities and UQ Bounds for Hypocoercive MCMC Samplers
• Non-reversible Markov chain Monte Carlo for sampling of districting maps
• Non-Convex Optimization via Non-Reversible Stochastic Gradient Langevin Dynamics
• Hessian-Free High-Resolution Nesterov Acceleration for Sampling
• Constructing sampling schemes via coupling: Markov semigroups and optimal transport
• Variational Formulas of Asymptotic Variance for General Discrete-time Markov Chains
• Long-time behaviour of degenerate diffusions: UFG-type SDEs and time-inhomogeneous hypoelliptic processes
• On the convergence of an improved and adaptive kinetic simulated annealing
• Reversible and non-reversible Markov Chain Monte Carlo algorithms for reservoir simulation problems
• Non-reversible sampling schemes on submanifolds
• Ensemble preconditioning for Markov chain Monte Carlo simulation
• Large Deviations of Irreversible Processes
• Pressure statistics from the path integral for Darcy flow through random porous media
• Generalizing Parallel Replica Dynamics: Trajectory Fragments, Asynchronous Computing, and PDMPs