High-dimensional scaling limits of piecewise deterministic sampling algorithms

DOI10.1214/21-AAP1762zbMATH Open1504.65002arXiv1807.11358MaRDI QIDQ2094570FDOQ2094570

Authors: Joris Bierkens, Kengo Kamatani, Gareth O. Roberts 

Publication date: 31 October 2022

Published in: The Annals of Applied Probability (Search for Journal in Brave)

Abstract: Piecewise deterministic Markov processes are an important new tool in the design of Markov Chain Monte Carlo algorithms. Two examples of fundamental importance are the Bouncy Particle Sampler (BPS) and the Zig-Zag process (ZZ). In this paper scaling limits for both algorithms are determined. Here the dimensionality of the space tends towards infinity and the target distribution is the multivariate standard normal distribution. For several quantities of interest (angular momentum, first coordinate, and negative log-density) the scaling limits show qualitatively very different and rich behaviour. Based on these scaling limits the performance of the two algorithms in high dimensions can be compared. Although for angular momentum both processes require only a computational effort of $O(d)$ to obtain approximately independent samples, the computational effort for negative log-density and first coordinate differ: for these BPS requires $O(d^2)$ computational effort whereas ZZ requires $O(d)$. Finally we provide a criterion for the choice of the refreshment rate of BPS.

Full work available at URL: https://arxiv.org/abs/1807.11358

Recommendations

• Adaptive schemes for piecewise deterministic Monte Carlo algorithms
• Randomized Hamiltonian Monte Carlo as scaling limit of the bouncy particle sampler and dimension-free convergence rates
• Hypocoercivity of piecewise deterministic Markov process-Monte Carlo
• Infinite dimensional piecewise deterministic Markov processes
• Limit theorems for the zig-zag process

zbMATH Keywords

Markov chain Monte CarloGaussian processweak convergenceexponential ergodicitypiecewise deterministic Markov processes

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Central limit and other weak theorems (60F05)

Cites Work

• Weak convergence and optimal scaling of random walk Metropolis algorithms
• Title not available (Why is that?)
• Title not available (Why is that?)
• Optimal scaling for various Metropolis-Hastings algorithms.
• Title not available (Why is that?)
• The zig-zag process and super-efficient sampling for Bayesian analysis of big data
• The Bouncy Particle Sampler: A Non-Reversible Rejection-Free Markov Chain Monte Carlo Method
• Coordinate sampler: a non-reversible Gibbs-like MCMC sampler
• Normal Approximation by Stein’s Method
• A dozen de Finetti-style results in search of a theory
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Normal approximations with Malliavin calculus. From Stein's method to universality
• Stability and Ergodicity of Piecewise Deterministic Markov Processes
• Title not available (Why is that?)
• Title not available (Why is that?)
• Exponential and uniform ergodicity of Markov processes
• Markov Processes, Gaussian Processes, and Local Times
• Equivalence of Stochastic Equations and Martingale Problems
• Long time behavior of telegraph processes under convex potentials
• Central limit theorems in D[0, 1]
• A piecewise deterministic Monte Carlo method for diffusion bridges
• Complexity bounds for Markov chain Monte Carlo algorithms via diffusion limits
• Hypocoercivity of piecewise deterministic Markov process-Monte Carlo
• Limit theorems for the zig-zag process
• Efficient strategy for the Markov chain Monte Carlo in high-dimension with heavy-tailed target probability distribution
• Fluctuations of the empirical measure of freezing Markov chains
• Exponential ergodicity of the bouncy particle sampler
• Ergodicity of the zigzag process

Cited In (13)

• On explicit \(L^2\)-convergence rate estimate for piecewise deterministic Markov processes in MCMC algorithms
• Infinite dimensional piecewise deterministic Markov processes
• On Unifying the Space of ℓ₀-Sampling Algorithms
• Polynomial convergence rates of piecewise deterministic Markov processes
• Optimal convergence rates of high order Parzen windows with unbounded sampling
• The Moser-Tardos Resample algorithm: Where is the limit? (an experimental inquiry)
• Adaptive schemes for piecewise deterministic Monte Carlo algorithms
• Posterior computation with the Gibbs zig-zag sampler
• Large deviations for the empirical measure of the zig-zag process
• Ergodicity of the zigzag process
• Sampling algorithms in statistical physics: a guide for statistics and machine learning
• Approximating Large Frequency Moments with Pick-and-Drop Sampling
• Limit theorems for the zig-zag process

This page was built for publication: High-dimensional scaling limits of piecewise deterministic sampling algorithms

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2094570)