Modified Cholesky Riemann manifold Hamiltonian Monte Carlo: exploiting sparsity for fast sampling of high-dimensional targets
From MaRDI portal
Publication:1704017
Abstract: Riemann manifold Hamiltonian Monte Carlo (RMHMC) has the potential to produce high-quality Markov chain Monte Carlo-output even for very challenging target distributions. To this end, a symmetric positive definite scaling matrix for RMHMC, which derives, via a modified Cholesky factorization, from the potentially indefinite negative Hessian of the target log-density is proposed. The methodology is able to exploit the sparsity of the Hessian, stemming from conditional independence modeling assumptions, and thus admit fast implementation of RMHMC even for high-dimensional target distributions. Moreover, the methodology can exploit log-concave conditional target densities, often encountered in Bayesian hierarchical models, for faster sampling and more straight forward tuning. The proposed methodology is compared to alternatives for some challenging targets, and is illustrated by applying a state space model to real data.
Recommendations
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- A general metric for Riemannian manifold Hamiltonian Monte Carlo
- Bayesian adaptively updated Hamiltonian Monte Carlo with an application to high-dimensional BEKK GARCH models
- MALA-within-Gibbs samplers for high-dimensional distributions with sparse conditional structure
- Proximal Markov chain Monte Carlo algorithms
Cites work
- A general metric for Riemannian manifold Hamiltonian Monte Carlo
- A New Modified Cholesky Factorization
- A Revised Modified Cholesky Factorization Algorithm
- A stochastic Newton MCMC method for large-scale statistical inverse problems with application to seismic inversion
- Adaptive proposal distribution for random walk Metropolis algorithm
- Adaptive step size selection for Hessian-based manifold Langevin samplers
- An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach
- Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations (with discussion)
- Automatic approximation of the marginal likelihood in non-Gaussian hierarchical models
- Bayesian data analysis.
- Bayesian interference based only on simulated likelihood particle filter analysis of dynamic economic models
- Bayesian learning for neural networks
- Direct Methods for Sparse Linear Systems
- Fast reverse-mode automatic differentiation using expression templates in C++
- Fast sampling of Gaussian Markov random fields
- Gaussian Markov Random Fields
- scientific article; zbMATH DE number 3793774 (Why is no real title available?)
- scientific article; zbMATH DE number 2117879 (Why is no real title available?)
- scientific article; zbMATH DE number 1456803 (Why is no real title available?)
- scientific article; zbMATH DE number 961607 (Why is no real title available?)
- Langevin diffusions and Metropolis-Hastings algorithms
- Langevin diffusions and the Metropolis-adjusted Langevin algorithm
- Likelihood analysis of non-Gaussian measurement time series
- MCMC using Hamiltonian dynamics
- Monte Carlo strategies in scientific computing
- Newton-type methods for unconstrained and linearly constrained optimization
- Note on the Sampling Distribution for the Metropolis-Hastings Algorithm
- Numerical Optimization
- On markov chain monte carlo methods for nonlinear and non-gaussian state-space models
- On some properties of Markov chain Monte Carlo simulation methods based on the particle filter
- On the application of improved symplectic integrators in Hamiltonian Monte Carlo
- Optimal tuning of the hybrid Monte Carlo algorithm
- Particle efficient importance sampling
- Particle Gibbs with ancestor sampling
- Particle Markov chain Monte Carlo for efficient numerical simulation
- Practical Markov Chain Monte Carlo
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Simulating Hamiltonian Dynamics
- Slice sampling. (With discussions and rejoinder)
- The geometric foundations of Hamiltonian Monte Carlo
- The Gibbs sampler with particle efficient importance sampling for state-space models
- The no-U-turn sampler: adaptively setting path lengths in Hamiltonian Monte Carlo
Cited in
(8)- Dynamically Rescaled Hamiltonian Monte Carlo for Bayesian Hierarchical Models
- Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
- Importance Sampling-Based Transport Map Hamiltonian Monte Carlo for Bayesian Hierarchical Models
- A general metric for Riemannian manifold Hamiltonian Monte Carlo
- Bayesian adaptively updated Hamiltonian Monte Carlo with an application to high-dimensional BEKK GARCH models
- Log-density gradient covariance and automatic metric tensors for Riemann manifold Monte Carlo methods
- Tuning diagonal scale matrices for HMC
- Unbiasing Hamiltonian Monte Carlo algorithms for a general Hamiltonian function
This page was built for publication: Modified Cholesky Riemann manifold Hamiltonian Monte Carlo: exploiting sparsity for fast sampling of high-dimensional targets
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1704017)
