Informed proposals for local MCMC in discrete spaces
From MaRDI portal
(Redirected from Publication:5130627)
Abstract: There is a lack of methodological results to design efficient Markov chain Monte Carlo (MCMC) algorithms for statistical models with discrete-valued high-dimensional parameters. Motivated by this consideration, we propose a simple framework for the design of informed MCMC proposals (i.e. Metropolis-Hastings proposal distributions that appropriately incorporate local information about the target) which is naturally applicable to both discrete and continuous spaces. We explicitly characterize the class of optimal proposal distributions under this framework, which we refer to as locally-balanced proposals, and prove their Peskun-optimality in high-dimensional regimes. The resulting algorithms are straightforward to implement in discrete spaces and provide orders of magnitude improvements in efficiency compared to alternative MCMC schemes, including discrete versions of Hamiltonian Monte Carlo. Simulations are performed with both simulated and real datasets, including a detailed application to Bayesian record linkage. A direct connection with gradient-based MCMC suggests that locally-balanced proposals may be seen as a natural way to extend the latter to discrete spaces.
Recommendations
- Efficient MCMC estimation of discrete distributions
- Informed reversible jump algorithms
- Adaptive random neighbourhood informed Markov chain Monte Carlo for high-dimensional Bayesian variable selection
- Efficient MCMC sampling in dynamic mixture models
- Efficient Construction of Reversible Jump Markov Chain Monte Carlo Proposal Distributions
Cites work
- A hierarchical Bayesian approach to record linkage and population size problems
- A note on Metropolis-Hastings kernels for general state spaces
- An external field prior for the hidden Potts model with application to cone-beam computed tomography
- Auxiliary Gradient-Based Sampling Algorithms
- Bayesian alignment using hierarchical models, with applications in protein bioinformatics
- EM, MCMC, and chain flipping for structure from motion with unknown correspondence
- Entity resolution with empirically motivated priors
- Equation of state calculations by fast computing machines
- Fast Langevin based algorithm for MCMC in high dimensions
- Handbook of Markov Chain Monte Carlo
- scientific article; zbMATH DE number 3297798 (Why is no real title available?)
- Markov Chain Monte Carlo Data Association for Multi-Target Tracking
- Markov chains and mixing times. With a chapter on ``Coupling from the past by James G. Propp and David B. Wilson.
- Monte Carlo sampling methods using Markov chains and their applications
- New methods for small area estimation with linkage uncertainty
- Optimal Scaling of Discrete Approximations to Langevin Diffusions
- Optimum Monte-Carlo sampling using Markov chains
- Random partition models and complementary clustering of Anglo-Saxon place-names
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Sequential Monte Carlo on large binary sampling spaces
- The Multiple-Try Method and Local Optimization in Metropolis Sampling
- Theoretical limits of microclustering for record linkage
- Weak convergence and optimal scaling of random walk Metropolis algorithms
Cited in
(39)- Uncertainty modelling and computational aspects of data association
- MALA with annealed proposals: a generalization of locally and globally balanced proposal distributions
- Adaptive random neighbourhood informed Markov chain Monte Carlo for high-dimensional Bayesian variable selection
- A prior for record linkage based on allelic partitions
- A multiple imputation procedure for record linkage and causal inference to estimate the effects of home-delivered meals
- Informed reversible jump algorithms
- The G-Wishart Weighted Proposal Algorithm: Efficient Posterior Computation for Gaussian Graphical Models
- Nonreversible Jump Algorithms for Bayesian Nested Model Selection
- d-blink: Distributed End-to-End Bayesian Entity Resolution
- Random Partition Models for Microclustering Tasks
- Markov Kernels Local Aggregation for Noise Vanishing Distribution Sampling
- Bayesian Semiparametric Longitudinal Drift-Diffusion Mixed Models for Tone Learning in Adults
- The taxicab sampler: MCMC for discrete spaces with application to tree models
- A Bayesian Nonparametric Test for Cross-Group Differences Relative to a Control
- Complexity analysis of Bayesian learning of high-dimensional DAG models and their equivalence classes
- A fresh Take on ‘Barker Dynamics’ for MCMC
- A new flexible Bayesian hypothesis test for multivariate data
- Unlabelled landmark matching via Bayesian data selection, and application to cell matching across imaging modalities
- Gibbs Sampling for Mixtures in Order of Appearance: The Ordered Allocation Sampler
- Sampling algorithms in statistical physics: a guide for statistics and machine learning
- Learning Block Structured Graphs in Gaussian Graphical Models
- An asymptotic Peskun ordering and its application to lifted samplers
- Detecting changes in the transmission rate of a stochastic epidemic model
- Fast Bayesian Record Linkage With Record-Specific Disagreement Parameters
- Dimension-free relaxation times of informed MCMC samplers on discrete spaces
- Analysis of Multiple-try Metropolis via Poincaré inequalities
- Generative Filtering for Recursive Bayesian Inference with Streaming Data
- Optimal design of the Barker proposal and other locally balanced Metropolis-Hastings algorithms
- Averaging polyhazard models using piecewise deterministic Monte Carlo with applications to data with long-term survivors
- Non-stationary phase of the Metropolis-adjusted Langevin algorithm with annealed proposals
- On the fundamental limitations of multi-proposal Markov chain Monte Carlo algorithms
- On mixing rates for Bayesian CART
- Fast Bayesian Record Linkage for Streaming Data Contexts
- Efficiency of reversible MCMC methods: elementary derivations and applications to composite methods
- Theoretical guarantees for lifted samplers
- Preconditioned discrete-HAMS: a second-order irreversible discrete sampler
- A hybrid Markov chain Monte Carlo approach for structural learning in Bayesian networks based on variable blocking
- Bayesian Weibull quantile regression via Barker’s MCMC proposal: application to cancer survival data
- Efficient MCMC estimation of discrete distributions
This page was built for publication: Informed proposals for local MCMC in discrete spaces
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5130627)
