Scalable Bayesian inference for the inverse temperature of a hidden Potts model
From MaRDI portal
Publication:2297228
Abstract: The inverse temperature parameter of the Potts model governs the strength of spatial cohesion and therefore has a major influence over the resulting model fit. A difficulty arises from the dependence of an intractable normalising constant on the value of this parameter and thus there is no closed-form solution for sampling from the posterior distribution directly. There are a variety of computational approaches for sampling from the posterior without evaluating the normalising constant, including the exchange algorithm and approximate Bayesian computation (ABC). A serious drawback of these algorithms is that they do not scale well for models with a large state space, such as images with a million or more pixels. We introduce a parametric surrogate model, which approximates the score function using an integral curve. Our surrogate model incorporates known properties of the likelihood, such as heteroskedasticity and critical temperature. We demonstrate this method using synthetic data as well as remotely-sensed imagery from the Landsat-8 satellite. We achieve up to a hundredfold improvement in the elapsed runtime, compared to the exchange algorithm or ABC. An open source implementation of our algorithm is available in the R package "bayesImageS."
Recommendations
- Pre-processing for approximate Bayesian computation in image analysis
- Efficient MCMC for Gibbs random fields using pre-computation
- Adaptive construction of surrogates for the Bayesian solution of inverse problems
- Scalable posterior approximations for large-scale Bayesian inverse problems via likelihood-informed parameter and state reduction
- Bayesian computation for statistical models with intractable normalizing constants
Cites work
- A Bayesian Reassessment of Nearest-Neighbor Classification
- A bounding chain for Swendsen-Wang
- A double Metropolis–Hastings sampler for spatial models with intractable normalizing constants
- ABC likelihood-free methods for model choice in Gibbs random fields
- Accelerating pseudo-marginal MCMC using Gaussian processes
- Adaptive optimal scaling of Metropolis-Hastings algorithms using the Robbins-Monro process
- An efficient Markov chain Monte Carlo method for distributions with intractable normalising constants
- Bayesian indirect inference using a parametric auxiliary model
- Bayesian mixture models in a longitudinal setting for analysing sheep CAT scan images
- Bayesian optimization for likelihood-free inference of simulator-based statistical models
- CHIRAL POTTS MODEL: CORNER TRANSFER MATRICES AND PARAMETRIZATIONS
- Diagnostic tools for approximate Bayesian computation using the coverage property
- Efficient MCMC for Gibbs random fields using pre-computation
- Efficient recursions for general factorisable models
- Exact sampling with coupled Markov chains and applications to statistical mechanics
- Gaussian process modelling in approximate Bayesian computation to estimate horizontal gene transfer in bacteria
- Getting It Right
- Hidden Markov Models and Disease Mapping
- scientific article; zbMATH DE number 3513115 (Why is no real title available?)
- scientific article; zbMATH DE number 1369837 (Why is no real title available?)
- scientific article; zbMATH DE number 1870396 (Why is no real title available?)
- scientific article; zbMATH DE number 3076589 (Why is no real title available?)
- Inference for Discrete Markov Fields: The Simplest Nontrivial Case
- Lazy ABC
- Markov Random Field Modeling in Image Analysis
- On Russian roulette estimates for Bayesian inference with doubly-intractable likelihoods
- Optimal Bayesian experimental design for models with intractable likelihoods using indirect inference applied to biological process models
- Perfect simulation
- Perfect slice samplers
- Pre-processing for approximate Bayesian computation in image analysis
- Precomputing strategy for Hamiltonian Monte Carlo method based on regularity in parameter space
- Proper likelihoods for Bayesian analysis
- RcppArmadillo: accelerating R with high-performance C++ linear algebra
- Recursive computing and simulation-free inference for general factorizable models
- Scalable Bayesian inference for the inverse temperature of a hidden Potts model
- Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images
- Variance bounding and geometric ergodicity of Markov chain Monte Carlo kernels for approximate Bayesian computation
Cited in
(11)- Efficient MCMC for Gibbs random fields using pre-computation
- Gibbs posterior convergence and the thermodynamic formalism
- The node-wise pseudo-marginal method: model selection with spatial dependence on latent graphs
- Scalable Bayesian inference for the inverse temperature of a hidden Potts model
- bayesImageS
- Improving detection of changepoints in short and noisy time series with local correlations: connecting the events in pixel neighbourhoods
- Neural networks for parameter estimation in intractable models
- A synthetic likelihood approach for intractable Markov random fields
- Generalized Bayesian Inference for Discrete Intractable Likelihood
- Warped gradient-enhanced Gaussian process surrogate models for exponential family likelihoods with intractable normalizing constants
- A general Bayesian functional spatial partitioning method for multiple region discovery applied to prostate cancer MRI
This page was built for publication: Scalable Bayesian inference for the inverse temperature of a hidden Potts model
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2297228)
