Variational Inference in Gaussian Processes via Probabilistic Point Assimilation

[edit]

Nathaniel J. King, IBM
Neil D. Lawrence, University of Sheffield

Related Material

Abstract

We introduce a novel variational approach for approximate inference in Gaussian process (GP) models. The key advantages of our approach are the ease with which different noise models can be incorporated and improved speed of convergence. We refer to the algorithm as probabilistic point assimilation (PPA). We introduce the algorithm firstly using the ‘weight space’ view and then through its Gaussian process formulation. We illustrate the approach on several benchmark data sets.


@TechReport{king-ppa05,
  title = 	 {Variational Inference in Gaussian Processes via Probabilistic Point Assimilation},
  author = 	 {Nathaniel J. King and Neil D. Lawrence},
  year = 	 {2005},
  institution = 	 {The University of Sheffield, Department of Computer Science},
  number =       {CS-05-06},
  month = 	 {00},
  edit = 	 {https://github.com/lawrennd//publications/edit/gh-pages/_posts/2005-01-01-king-ppa05.md},
  url =  	 {http://inverseprobability.com/publications/king-ppa05.html},
  abstract = 	 {We introduce a novel variational approach for approximate inference in Gaussian process (GP) models. The key advantages of our approach are the ease with which different noise models can be incorporated and improved speed of convergence. We refer to the algorithm as probabilistic point assimilation (PPA). We introduce the algorithm firstly using the ‘weight space’ view and then through its Gaussian process formulation. We illustrate the approach on several benchmark data sets.},
  key = 	 {King:ppa05},
  linkpsgz =  {ftp://ftp.dcs.shef.ac.uk/home/neil/ppatech.ps.gz},
  group = 	 {shefml}
 

}
%T Variational Inference in Gaussian Processes via Probabilistic Point Assimilation
%A Nathaniel J. King and Neil D. Lawrence
%B 
%D 
%F king-ppa05	
%P --
%R 
%U http://inverseprobability.com/publications/king-ppa05.html
%N CS-05-06
%X We introduce a novel variational approach for approximate inference in Gaussian process (GP) models. The key advantages of our approach are the ease with which different noise models can be incorporated and improved speed of convergence. We refer to the algorithm as probabilistic point assimilation (PPA). We introduce the algorithm firstly using the ‘weight space’ view and then through its Gaussian process formulation. We illustrate the approach on several benchmark data sets.
TY  - CPAPER
TI  - Variational Inference in Gaussian Processes via Probabilistic Point Assimilation
AU  - Nathaniel J. King
AU  - Neil D. Lawrence
PY  - 2005/01/01
DA  - 2005/01/01	
ID  - king-ppa05	
SP  - 
EP  - 
UR  - http://inverseprobability.com/publications/king-ppa05.html
AB  - We introduce a novel variational approach for approximate inference in Gaussian process (GP) models. The key advantages of our approach are the ease with which different noise models can be incorporated and improved speed of convergence. We refer to the algorithm as probabilistic point assimilation (PPA). We introduce the algorithm firstly using the ‘weight space’ view and then through its Gaussian process formulation. We illustrate the approach on several benchmark data sets.
ER  -

King, N.J. & Lawrence, N.D.. (2005). Variational Inference in Gaussian Processes via Probabilistic Point Assimilation.(CS-05-06):-