Constrained ensemble Langevin Monte Carlo

DOI10.3934/FODS.2021034zbMATH Open1489.65008arXiv2102.04279OpenAlexW4205796992MaRDI QIDQ2148951FDOQ2148951

Qin Li, Zhiyan Ding

Publication date: 24 June 2022

Published in: Foundations of Data Science (Search for Journal in Brave)

Abstract: The classical Langevin Monte Carlo method looks for samples from a target distribution by descending the samples along the gradient of the target distribution. The method enjoys a fast convergence rate. However, the numerical cost is sometimes high because each iteration requires the computation of a gradient. One approach to eliminate the gradient computation is to employ the concept of ``ensemble." A large number of particles are evolved together so the neighboring particles provide gradient information to each other. In this article, we discuss two algorithms that integrate the ensemble feature into LMC and the associated properties. In particular, we find that if one directly surrogates the gradient using the ensemble approximation, the algorithm, termed Ensemble Langevin Monte Carlo, is unstable due to a high variance term. If the gradients are replaced by the ensemble approximations only in a constrained manner, to protect from the unstable points, the algorithm, termed Constrained Ensemble Langevin Monte Carlo, resembles the classical LMC up to an ensemble error but removes most of the gradient computation.

Full work available at URL: https://arxiv.org/abs/2102.04279

Recommendations

• Higher order Langevin Monte Carlo algorithm
• Particle Metropolis-adjusted Langevin algorithms
• Analysis of Langevin Monte Carlo via convex optimization
• Sampling from a log-concave distribution with projected Langevin Monte Carlo
• User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient

zbMATH Keywords

varianceensemble methodsLangevin Monte Carlogradient free

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Stochastic methods (Fokker-Planck, Langevin, etc.) applied to problems in time-dependent statistical mechanics (82C31)

Cites Work

• Ensemble preconditioning for Markov chain Monte Carlo simulation
• Exponential convergence of Langevin distributions and their discrete approximations
• Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images
• Monte Carlo sampling methods using Markov chains and their applications
• Langevin diffusions and Metropolis-Hastings algorithms
• General state space Markov chains and MCMC algorithms
• On the trend to equilibrium for the Fokker-Planck equation: an interplay between physics and functional analysis.
• A stochastic Newton MCMC method for large-scale statistical inverse problems with application to seismic inversion
• An introduction to MCMC for machine learning
• The ensemble Kalman filter for combined state and parameter estimation
• Title not available (Why is that?)
• A dynamical systems framework for intermittent data assimilation
• Ensemble Kalman methods for inverse problems
• Multilevel sequential Monte Carlo samplers
• Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
• Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
• Ensemble Kalman Sampler: Mean-field Limit and Convergence Analysis
• Analysis of the Ensemble Kalman Filter for Inverse Problems
• On sampling from a log-concave density using kinetic Langevin diffusions
• Title not available (Why is that?)
• Title not available (Why is that?)
• Ensemble Kalman inversion: mean-field limit and convergence analysis
• Interacting Langevin Diffusions: Gradient Structure and Ensemble Kalman Sampler
• User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient
• MALA-within-Gibbs Samplers for High-Dimensional Distributions with Sparse Conditional Structure
• Affine Invariant Interacting Langevin Dynamics for Bayesian Inference
• Continuous limits for constrained ensemble Kalman filter
• Title not available (Why is that?)

Cited In (3)

• Ensemble-Based Gradient Inference for Particle Methods in Optimization and Sampling
• Extended ensemble Monte Carlo approach to hardly relaxing problems
• CPMC-lab: a Matlab package for constrained path Monte Carlo calculations

Uses Software

• EnKF
• EnsembleQN

This page was built for publication: Constrained ensemble Langevin Monte Carlo

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2148951)