A MCMC method based on surrogate model and Gaussian process parameterization for infinite Bayesian PDE inversion.
Zheng HuHongqiao WangQingping ZhouPublished in: J. Comput. Phys. (2024)
Keyphrases
- gaussian process
- fully bayesian
- expectation propagation
- bayesian inference
- probabilistic model
- gaussian processes
- bayesian framework
- dynamical model
- hyperparameters
- markov chain monte carlo
- posterior distribution
- similarity measure
- marginal likelihood
- em algorithm
- gaussian process regression
- closed form
- prior information
- gaussian process models
- prior knowledge
- image processing
- sparse approximation
- model selection
- parameter estimation
- energy function
- cross validation
- bayesian methods
- markov chain
- variational inference
- maximum likelihood
- pairwise
- covariance function
- gaussian distribution
- posterior probability