The Split Gibbs Sampler Revisited: Improvements to Its Algorithmic Structure and Augmented Target Distribution
From MaRDI portal
Publication:6144058
Abstract: Developing efficient Bayesian computation algorithms for imaging inverse problems is challenging due to the dimensionality involved and because Bayesian imaging models are often not smooth. Current state-of-the-art methods often address these difficulties by replacing the posterior density with a smooth approximation that is amenable to efficient exploration by using Langevin Markov chain Monte Carlo (MCMC) methods. An alternative approach is based on data augmentation and relaxation, where auxiliary variables are introduced in order to construct an approximate augmented posterior distribution that is amenable to efficient exploration by Gibbs sampling. This paper proposes a new accelerated proximal MCMC method called latent space SK-ROCK (ls SK-ROCK), which tightly combines the benefits of the two aforementioned strategies. Additionally, instead of viewing the augmented posterior distribution as an approximation of the original model, we propose to consider it as a generalisation of this model. Following on from this, we empirically show that there is a range of values for the relaxation parameter for which the accuracy of the model improves, and propose a stochastic optimisation algorithm to automatically identify the optimal amount of relaxation for a given problem. In this regime, ls SK-ROCK converges faster than competing approaches from the state of the art, and also achieves better accuracy since the underlying augmented Bayesian model has a higher Bayesian evidence. The proposed methodology is demonstrated with a range of numerical experiments related to image deblurring and inpainting, as well as with comparisons with alternative approaches from the state of the art. An open-source implementation of the proposed MCMC methods is available from https://github.com/luisvargasmieles/ls-MCMC.
Recommendations
- Accelerated Bayesian imaging by relaxed proximal-point Langevin sampling
- Efficient Bayesian computation by proximal Markov chain Monte Carlo: when Langevin meets Moreau
- A proximal Markov chain Monte Carlo method for Bayesian inference in imaging inverse problems: when Langevin meets Moreau
- Fast Gibbs sampling for high-dimensional Bayesian inversion
- Efficient Bayesian Computation for Low-Photon Imaging Problems
Cites work
- A nonlocal Bayesian image denoising algorithm
- Accelerating Proximal Markov Chain Monte Carlo by Using an Explicit Stabilized Method
- An Augmented Lagrangian Approach to the Constrained Optimization Formulation of Imaging Inverse Problems
- An introduction to continuous optimization for imaging
- An introduction to MCMC for machine learning
- Bayesian Compressive Sensing Using Laplace Priors
- Bayesian Imaging Using Plug & Play Priors: When Langevin Meets Tweedie
- Distributed optimization and statistical learning via the alternating direction method of multipliers
- Efficient Bayesian computation by proximal Markov chain Monte Carlo: when Langevin meets Moreau
- Efficient Gaussian Sampling for Solving Large-Scale Inverse Problems Using MCMC
- Efficient stochastic optimisation by unadjusted Langevin Monte Carlo. Application to maximum marginal likelihood and empirical Bayesian estimation
- Exponential convergence of Langevin distributions and their discrete approximations
- Fast Image Recovery Using Variable Splitting and Constrained Optimization
- Global Consensus Monte Carlo
- High-dimensional Bayesian inference via the unadjusted Langevin algorithm
- High-dimensional Gaussian sampling: a review and a unifying approach based on a stochastic proximal point algorithm
- scientific article; zbMATH DE number 3567782 (Why is no real title available?)
- scientific article; zbMATH DE number 2117879 (Why is no real title available?)
- scientific article; zbMATH DE number 7625178 (Why is no real title available?)
- scientific article; zbMATH DE number 3293978 (Why is no real title available?)
- Image Deblurring and Super-Resolution by Adaptive Sparse Domain Selection and Adaptive Regularization
- Image recovery via total variation minimization and related problems
- Maximum Likelihood Estimation of Misspecified Models
- Maximum likelihood estimation of regularization parameters in high-dimensional inverse problems: an empirical Bayesian approach. I: Methodology and experiments
- Maximum likelihood estimation of regularization parameters in high-dimensional inverse problems: an empirical Bayesian approach. II: Theoretical analysis
- Maximum-a-posteriori estimation with Bayesian confidence regions
- Nonasymptotic convergence analysis for the unadjusted Langevin algorithm
- Nonlinear time series. Theory, methods and applications with R examples
- Optimal explicit stabilized integrator of weak order 1 for stiff and ergodic stochastic differential equations
- Patch-based image restoration using expectation propagation
- Posterior expectation of the total variation model: properties and experiments
- Proximal Markov chain Monte Carlo algorithms
- Proximal nested sampling for high-dimensional Bayesian model selection
- Scalable Bayesian uncertainty quantification in imaging inverse problems via convex optimization
- Solving inverse problems using data-driven models
- Split-and-Augmented Gibbs Sampler—Application to Large-Scale Inference Problems
- Statistical and computational inverse problems.
- Stochastic proximal-gradient algorithms for penalized mixed models
- Template-based image reconstruction from sparse tomographic data
- The Bayesian Choice
- The Calculation of Posterior Distributions by Data Augmentation
- Total generalized variation
- Variational methods in imaging
Cited in
(5)- Properties of the bridge sampler with a focus on splitting the MCMC sample
- Accelerated Bayesian imaging by relaxed proximal-point Langevin sampling
- Marginal likelihood estimation in semiblind image deconvolution: a stochastic approximation approach
- Empirical Bayesian image restoration by Langevin sampling with a denoising diffusion implicit prior
- Sampling strategies in Bayesian inversion: a study of RTO and Langevin methods
This page was built for publication: The Split Gibbs Sampler Revisited: Improvements to Its Algorithmic Structure and Augmented Target Distribution
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6144058)
