Piecewise deterministic Markov processes for continuous-time Monte Carlo

DOI10.1214/18-STS648zbMATH Open1403.62148arXiv1611.07873OpenAlexW2951601728MaRDI QIDQ1630397FDOQ1630397

Authors: Paul Fearnhead, Joris Bierkens, Murray Pollock, Gareth O. Roberts 

Publication date: 10 December 2018

Published in: Statistical Science (Search for Journal in Brave)

Abstract: Recently there have been exciting developments in Monte Carlo methods, with the development of new MCMC and sequential Monte Carlo (SMC) algorithms which are based on continuous-time, rather than discrete-time, Markov processes. This has led to some fundamentally new Monte Carlo algorithms which can be used to sample from, say, a posterior distribution. Interestingly, continuous-time algorithms seem particularly well suited to Bayesian analysis in big-data settings as they need only access a small sub-set of data points at each iteration, and yet are still guaranteed to target the true posterior distribution. Whilst continuous-time MCMC and SMC methods have been developed independently we show here that they are related by the fact that both involve simulating a piecewise deterministic Markov process. Furthermore we show that the methods developed to date are just specific cases of a potentially much wider class of continuous-time Monte Carlo algorithms. We give an informal introduction to piecewise deterministic Markov processes, covering the aspects relevant to these new Monte Carlo algorithms, with a view to making the development of new continuous-time Monte Carlo more accessible. We focus on how and why sub-sampling ideas can be used with these algorithms, and aim to give insight into how these new algorithms can be implemented, and what are some of the issues that affect their efficiency.

Full work available at URL: https://arxiv.org/abs/1611.07873

Recommendations

• Piecewise deterministic Markov processes for scalable Monte Carlo on restricted domains
• Piecewise-deterministic Markov processes
• Piecewise deterministic Markov processes
• Hypocoercivity of piecewise deterministic Markov process-Monte Carlo
• Piecewise deterministic Markov process -- recent results
• scientific article; zbMATH DE number 3901778
• Statistical inference for piecewise-deterministic Markov processes
• A class of piecewise deterministic Markov processes

zbMATH Keywords

Bayesian statisticscontrol variatesbig databouncy particle samplercontinuous-time importance samplingSCALEzig-zag sampler

Mathematics Subject Classification ID

Bayesian inference (62F15) Markov processes: hypothesis testing (62M02)

Cites Work

• MCMC using Hamiltonian dynamics
• Simulation of nonhomogeneous poisson processes by thinning
• Slice sampling. (With discussions and rejoinder)
• The pseudo-marginal approach for efficient Monte Carlo computations
• Analysis of a nonreversible Markov chain sampler.
• Title not available (Why is that?)
• Blind Deconvolution via Sequential Imputations
• On the stability of interacting processes with applications to filtering and genetic algorithms
• Sequential Monte Carlo Methods for Dynamic Systems
• Title not available (Why is that?)
• Retrospective exact simulation of diffusion sample paths with applications
• Title not available (Why is that?)
• Optimal Scaling of Discrete Approximations to Langevin Diffusions
• The Bouncy Particle Sampler: A Non-Reversible Rejection-Free Markov Chain Monte Carlo Method
• Title not available (Why is that?)
• Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
• Title not available (Why is that?)
• Speeding Up MCMC by Efficient Data Subsampling
• Title not available (Why is that?)
• Stochastic differential equations. An introduction with applications
• Simulation of Brownian motion at first-passage times
• Exact simulation of diffusions
• On the exact and \(\varepsilon\)-strong simulation of (jump) diffusions
• A factorisation of diffusion measure and finite sample path constructions
• Title not available (Why is that?)
• Long-term stability of sequential Monte Carlo methods under verifiable conditions
• Non-reversible Metropolis-Hastings
• On Russian roulette estimates for Bayesian inference with doubly-intractable likelihoods
• Particle Filters for Partially Observed Diffusions
• A short history of Markov chain Monte Carlo: Subjective recollections from incomplete data
• A piecewise deterministic scaling limit of lifted Metropolis-Hastings in the Curie-Weiss model
• Simple, scalable and accurate posterior interval estimation
• Piecewise deterministic Markov processes for scalable Monte Carlo on restricted domains
• Limit theorems for the zig-zag process

Cited In (43)

• Exponential ergodicity of a degenerate age-size piecewise deterministic process
• On explicit \(L^2\)-convergence rate estimate for piecewise deterministic Markov processes in MCMC algorithms
• Speed up Zig-Zag
• Automatic zig-zag sampling in practice
• Collective proposal distributions for nonlinear MCMC samplers: mean-field theory and fast implementation
• Informed sub-sampling MCMC: approximate Bayesian inference for large datasets
• Transport Monte Carlo: High-Accuracy Posterior Approximation via Random Transport
• Analysis of stochastic gradient descent in continuous time
• Reversible Jump PDMP Samplers for Variable Selection
• Exponential ergodicity for damping Hamiltonian dynamics with state-dependent and non-local collisions
• Piecewise deterministic Markov processes for scalable Monte Carlo on restricted domains
• Non-reversible processes: GENERIC, hypocoercivity and fluctuations
• Constructing sampling schemes via coupling: Markov semigroups and optimal transport
• Coordinate sampler: a non-reversible Gibbs-like MCMC sampler
• Low-lying eigenvalues and convergence to the equilibrium of some piecewise deterministic Markov processes generators in the small temperature regime
• Log-density gradient covariance and automatic metric tensors for Riemann manifold Monte Carlo methods
• Quantifying and Managing Uncertainty in Piecewise-Deterministic Markov Processes
• Concave-Convex PDMP-based Sampling
• Stochastic gradient Markov chain Monte Carlo
• Speeding up the zig-zag process
• Subgeometric hypocoercivity for piecewise-deterministic Markov process Monte Carlo methods
• Highly Scalable Bayesian Geostatistical Modeling via Meshed Gaussian Processes on Partitioned Domains
• Generalized Bayesian likelihood-free inference
• An MCMC computational approach for a continuous time state-dependent regime switching diffusion process
• Adaptive schemes for piecewise deterministic Monte Carlo algorithms
• Strong invariance principles for ergodic Markov processes
• Peskun-Tierney ordering for Markovian Monte Carlo: beyond the reversible scenario
• Tuning diagonal scale matrices for HMC
• Posterior computation with the Gibbs zig-zag sampler
• A short introduction to piecewise deterministic Markov samplers
• A note on the polynomial ergodicity of the one-dimensional Zig-Zag process
• A piecewise deterministic Monte Carlo method for diffusion bridges
• Connecting the Dots: Numerical Randomized Hamiltonian Monte Carlo with State-Dependent Event Rates
• Large deviations for the empirical measure of the zig-zag process
• PDMP characterisation of event-chain Monte Carlo algorithms for particle systems
• A Quasi Monte Carlo Approach to Piecewise Linear Markov Approximations of Markov Operators
• Title not available (Why is that?)
• The zig-zag process and super-efficient sampling for Bayesian analysis of big data
• Cores for piecewise-deterministic Markov processes used in Markov chain Monte Carlo
• Computing Bayes: from then `til now
• Editorial: Bayesian computations in the 21st century
• Sampling algorithms in statistical physics: a guide for statistics and machine learning
• Approximations of piecewise deterministic Markov processes and their convergence properties

This page was built for publication: Piecewise deterministic Markov processes for continuous-time Monte Carlo

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1630397)