Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4)

Jarno Vanhatalo, Jaakko Riihimäki, Jouni Hartikainen, Pasi Jylänki, Ville Tolvanen, Aki Vehtari

Research output: Book/ReportCommissioned reportProfessional

Abstract

Gaussian processes (GP) are powerful tools for probabilistic modeling purposes. They can be used to define prior distributions over latent functions in hierarchical Bayesian models. The prior over functions is defined implicitly by the mean and covariance function, which determine the smoothness and variability of the function. The inference can then be conducted directly in the function space by evaluating or approximating the posterior process. Despite their attractive theoretical properties GPs provide practical challenges in their implementation. GPstuff is a versatile collection of computational tools for GP models. It has been implemented for Linux and Windows MATLAB and includes, among others, various inference methods, sparse approximations and tools for model assessment. In this work, we review these tools and demonstrate the use of GPstuff in several models.
Original languageEnglish
Number of pages58
Publication statusPublished - 5 Oct 2012
MoE publication typeD4 Published development or research report or study

Fields of Science

  • 113 Computer and information sciences

Cite this

Vanhatalo, J., Riihimäki, J., Hartikainen, J., Jylänki, P., Tolvanen, V., & Vehtari, A. (2012). Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4).
Vanhatalo, Jarno ; Riihimäki, Jaakko ; Hartikainen, Jouni ; Jylänki, Pasi ; Tolvanen, Ville ; Vehtari, Aki. / Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4). 2012. 58 p.
@book{8a2ca550f35f431087c7489d8ba29852,
title = "Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4)",
abstract = "Gaussian processes (GP) are powerful tools for probabilistic modeling purposes. They can be used to define prior distributions over latent functions in hierarchical Bayesian models. The prior over functions is defined implicitly by the mean and covariance function, which determine the smoothness and variability of the function. The inference can then be conducted directly in the function space by evaluating or approximating the posterior process. Despite their attractive theoretical properties GPs provide practical challenges in their implementation. GPstuff is a versatile collection of computational tools for GP models. It has been implemented for Linux and Windows MATLAB and includes, among others, various inference methods, sparse approximations and tools for model assessment. In this work, we review these tools and demonstrate the use of GPstuff in several models.",
keywords = "113 Computer and information sciences",
author = "Jarno Vanhatalo and Jaakko Riihim{\"a}ki and Jouni Hartikainen and Pasi Jyl{\"a}nki and Ville Tolvanen and Aki Vehtari",
year = "2012",
month = "10",
day = "5",
language = "English",

}

Vanhatalo, J, Riihimäki, J, Hartikainen, J, Jylänki, P, Tolvanen, V & Vehtari, A 2012, Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4).

Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4). / Vanhatalo, Jarno; Riihimäki, Jaakko; Hartikainen, Jouni; Jylänki, Pasi; Tolvanen, Ville; Vehtari, Aki.

2012. 58 p.

Research output: Book/ReportCommissioned reportProfessional

TY - BOOK

T1 - Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4)

AU - Vanhatalo, Jarno

AU - Riihimäki, Jaakko

AU - Hartikainen, Jouni

AU - Jylänki, Pasi

AU - Tolvanen, Ville

AU - Vehtari, Aki

PY - 2012/10/5

Y1 - 2012/10/5

N2 - Gaussian processes (GP) are powerful tools for probabilistic modeling purposes. They can be used to define prior distributions over latent functions in hierarchical Bayesian models. The prior over functions is defined implicitly by the mean and covariance function, which determine the smoothness and variability of the function. The inference can then be conducted directly in the function space by evaluating or approximating the posterior process. Despite their attractive theoretical properties GPs provide practical challenges in their implementation. GPstuff is a versatile collection of computational tools for GP models. It has been implemented for Linux and Windows MATLAB and includes, among others, various inference methods, sparse approximations and tools for model assessment. In this work, we review these tools and demonstrate the use of GPstuff in several models.

AB - Gaussian processes (GP) are powerful tools for probabilistic modeling purposes. They can be used to define prior distributions over latent functions in hierarchical Bayesian models. The prior over functions is defined implicitly by the mean and covariance function, which determine the smoothness and variability of the function. The inference can then be conducted directly in the function space by evaluating or approximating the posterior process. Despite their attractive theoretical properties GPs provide practical challenges in their implementation. GPstuff is a versatile collection of computational tools for GP models. It has been implemented for Linux and Windows MATLAB and includes, among others, various inference methods, sparse approximations and tools for model assessment. In this work, we review these tools and demonstrate the use of GPstuff in several models.

KW - 113 Computer and information sciences

M3 - Commissioned report

BT - Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4)

ER -

Vanhatalo J, Riihimäki J, Hartikainen J, Jylänki P, Tolvanen V, Vehtari A. Bayesian Modeling with Gaussian Processes using the MATLAB Toolbox GPstuff (v3.4). 2012. 58 p.