CUED Publications database

Deep Gaussian Processes with Decoupled Inducing Inputs

Havasi, M and Hernández-Lobato, JM and Murillo-Fuentes, JJ Deep Gaussian Processes with Decoupled Inducing Inputs. (Unpublished)

Full text not available from this repository.

Abstract

Deep Gaussian Processes (DGP) are hierarchical generalizations of Gaussian Processes (GP) that have proven to work effectively on a multiple supervised regression tasks. They combine the well calibrated uncertainty estimates of GPs with the great flexibility of multilayer models. In DGPs, given the inputs, the outputs of the layers are Gaussian distributions parameterized by their means and covariances. These layers are realized as Sparse GPs where the training data is approximated using a small set of pseudo points. In this work, we show that the computational cost of DGPs can be reduced with no loss in performance by using a separate, smaller set of pseudo points when calculating the layerwise variance while using a larger set of pseudo points when calculating the layerwise mean. This enabled us to train larger models that have lower cost and better predictive performance.

Item Type: Article
Uncontrolled Keywords: stat.ML stat.ML
Subjects: UNSPECIFIED
Divisions: Div F > Computational and Biological Learning
Depositing User: Cron Job
Date Deposited: 10 Jan 2018 20:07
Last Modified: 18 Aug 2020 12:42
DOI: