A function space HMC algorithm with second order Langevin diffusion limit

DOI10.3150/14-BEJ621zbMATH Open1346.60119arXiv1308.0543OpenAlexW1872439012MaRDI QIDQ5963495FDOQ5963495

A. M. Stuart, Natesh S. Pillai, M. Ottobre, Frank J. Pinski

Publication date: 22 February 2016

Published in: Bernoulli (Search for Journal in Brave)

Abstract: We describe a new MCMC method optimized for the sampling of probability measures on Hilbert space which have a density with respect to a Gaussian; such measures arise in the Bayesian approach to inverse problems, and in conditioned diffusions. Our algorithm is based on two key design principles: (i) algorithms which are well-defined in infinite dimensions result in methods which do not suffer from the curse of dimensionality when they are applied to approximations of the infinite dimensional target measure on $$; (ii) non-reversible algorithms can have better mixing properties compared to their reversible counterparts. The method we introduce is based on the hybrid Monte Carlo algorithm, tailored to incorporate these two design principles. The main result of this paper states that the new algorithm, appropriately rescaled, converges weakly to a second order Langevin diffusion on Hilbert space; as a consequence the algorithm explores the approximate target measures on $$ in a number of steps which is independent of $N$. We also present the underlying theory for the limiting non-reversible diffusion on Hilbert space, including characterization of the invariant measure, and we describe numerical simulations demonstrating that the proposed method has favourable mixing properties as an MCMC algorithm.

Full work available at URL: https://arxiv.org/abs/1308.0543

Recommendations

• Hybrid Monte Carlo on Hilbert spaces
• Spectral gaps for a Metropolis-Hastings algorithm in infinite dimensions
• Mixing rates for Hamiltonian Monte Carlo algorithms in finite and infinite dimensions
• Diffusion limits of the random walk Metropolis algorithm in high dimensions
• Optimal scaling and diffusion limits for the Langevin algorithm in high dimensions

zbMATH Keywords

diffusion limitsMarkov chain Monte Carlo algorithmhybrid Monte Carlo algorithmsecond-order Langevin diffusion

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Monte Carlo methods (65C05) Numerical analysis or methods applied to Markov chains (65C40) Central limit and other weak theorems (60F05) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Diffusion processes (60J60)

Cites Work

• Title not available (Why is that?)
• Analysis of a nonreversible Markov chain sampler.
• Monte Carlo sampling methods using Markov chains and their applications
• Inverse problems: a Bayesian perspective
• Title not available (Why is that?)
• A concise course on stochastic partial differential equations
• Invariant measure of the stochastic Allen-Cahn equation: the regime of small noise and large system size
• Accelerating diffusions
• MCMC methods for functions: modifying old algorithms to make them faster
• Accelerating Gaussian diffusions
• MCMC METHODS FOR DIFFUSION BRIDGES
• Hybrid Monte Carlo on Hilbert spaces
• A patch that imparts unconditional stability to explicit integrators for Langevin-like equations
• Asymptotic behaviour of a class of stochastic approximation procedures
• Pathwise accuracy and ergodicity of metropolized integrators for SDEs
• Title not available (Why is that?)
• Diffusion limits of the random walk Metropolis algorithm in high dimensions
• Optimal scaling and diffusion limits for the Langevin algorithm in high dimensions
• Noisy gradient flow from a random walk in Hilbert space

Cited In (26)

• High-dimensional MCMC with a standard splitting scheme for the underdamped Langevin diffusion
• On the geometric ergodicity of Hamiltonian Monte Carlo
• Continuum limit and preconditioned Langevin sampling of the path integral molecular dynamics
• Probability and moment inequalities for additive functionals of geometrically ergodic Markov chains
• Reversible and non-reversible Markov chain Monte Carlo algorithms for reservoir simulation problems
• Non-reversible processes: GENERIC, hypocoercivity and fluctuations
• Exact targeting of Gibbs distributions using velocity-jump processes
• Using perturbed underdamped Langevin dynamics to efficiently sample from probability distributions
• Modified Hamiltonian Monte Carlo for Bayesian inference
• Two Metropolis--Hastings Algorithms for Posterior Measures with Non-Gaussian Priors in Infinite Dimensions
• Geometric Integration of Measure-Preserving Flows for Sampling
• Is there an analog of Nesterov acceleration for gradient-based MCMC?
• FEM-based discretization-invariant MCMC methods for PDE-constrained Bayesian inverse problems
• Irreversible samplers from jump and continuous Markov processes
• On Irreversible Metropolis Sampling Related to Langevin Dynamics
• Diffusion limit for the random walk Metropolis algorithm out of stationarity
• Mixing rates for Hamiltonian Monte Carlo algorithms in finite and infinite dimensions
• Peskun-Tierney ordering for Markovian Monte Carlo: beyond the reversible scenario
• Non-reversible guided Metropolis kernel
• Hamiltonian-Assisted Metropolis Sampling
• Advanced Multilevel Monte Carlo Methods
• Markov chain Monte Carlo and irreversibility
• Optimal scaling of the MALA algorithm with irreversible proposals for Gaussian targets
• MALA-within-Gibbs Samplers for High-Dimensional Distributions with Sparse Conditional Structure
• Limit theorems for the zig-zag process
• Analysis of a Computational Framework for Bayesian Inverse Problems: Ensemble Kalman Updates and MAP Estimators under Mesh Refinement

This page was built for publication: A function space HMC algorithm with second order Langevin diffusion limit

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5963495)