Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Deep Gaussian Process Emulation using Stochastic Imputation
(Submitted on 4 Jul 2021 (v1), last revised 17 Aug 2022 (this version, v2))
Abstract: Deep Gaussian processes (DGPs) provide a rich class of models that can better represent functions with varying regimes or sharp changes, compared to conventional GPs. In this work, we propose a novel inference method for DGPs for computer model emulation. By stochastically imputing the latent layers, our approach transforms a DGP into a linked GP: a novel emulator developed for systems of linked computer models. This transformation permits an efficient DGP training procedure that only involves optimizations of conventional GPs. In addition, predictions from DGP emulators can be made in a fast and analytically tractable manner by naturally utilizing the closed form predictive means and variances of linked GP emulators. We demonstrate the method in a series of synthetic examples and empirical applications, and show that it is a competitive candidate for DGP surrogate inference, combining efficiency that is comparable to doubly stochastic variational inference and uncertainty quantification that is comparable to the fully-Bayesian approach. A $\texttt{Python}$ package $\texttt{dgpsi}$ implementing the method is also produced and available at this https URL
Submission history
From: Deyu Ming [view email][v1] Sun, 4 Jul 2021 10:46:23 GMT (14044kb,D)
[v2] Wed, 17 Aug 2022 11:46:51 GMT (16045kb,D)
Link back to: arXiv, form interface, contact.