Semi-supervised Learning via Gaussian Processes

[edit]

Neil D. Lawrence, University of Sheffield
Michael I. Jordan, UC Berkeley

in Advances in Neural Information Processing Systems 17, pp 753-760

Related Material

Abstract

We present a probabilistic approach to learning a Gaussian Process classifier in the presence of unlabeled data. Our approach involves a “null category noise model” (NCNM) inspired by ordered categorical noise models. The noise model reflects an assumption that the data density is lower between the class-conditional densities. We illustrate our approach on a toy problem and present comparative results for the semi-supervised classification of handwritten digits.


@InProceedings{lawrence-semisuper04,
  title = 	 {Semi-supervised Learning via Gaussian Processes},
  author = 	 {Neil D. Lawrence and Michael I. Jordan},
  booktitle = 	 {Advances in Neural Information Processing Systems},
  pages = 	 {753},
  year = 	 {2005},
  editor = 	 {Lawrence Saul and Yair Weiss and Léon Bouttou},
  volume = 	 {17},
  address = 	 {Cambridge, MA},
  month = 	 {00},
  publisher = 	 {MIT Press},
  edit = 	 {https://github.com/lawrennd//publications/edit/gh-pages/_posts/2005-01-01-lawrence-semisuper04.md},
  url =  	 {http://inverseprobability.com/publications/lawrence-semisuper04.html},
  abstract = 	 {We present a probabilistic approach to learning a Gaussian Process classifier in the presence of unlabeled data. Our approach involves a “null category noise model” (NCNM) inspired by ordered categorical noise models. The noise model reflects an assumption that the data density is lower between the class-conditional densities. We illustrate our approach on a toy problem and present comparative results for the semi-supervised classification of handwritten digits.},
  crossref =  {Saul:nips04},
  key = 	 {Lawrence:semisuper04},
  linkpsgz =  {ftp://ftp.dcs.shef.ac.uk/home/neil/ncnm.ps.gz},
  linksoftware = {http://inverseprobability.com/ncnm/},
  group = 	 {shefml}
 

}
%T Semi-supervised Learning via Gaussian Processes
%A Neil D. Lawrence and Michael I. Jordan
%B 
%C Advances in Neural Information Processing Systems
%D 
%E Lawrence Saul and Yair Weiss and Léon Bouttou
%F lawrence-semisuper04
%I MIT Press	
%P 753--760
%R 
%U http://inverseprobability.com/publications/lawrence-semisuper04.html
%V 17
%X We present a probabilistic approach to learning a Gaussian Process classifier in the presence of unlabeled data. Our approach involves a “null category noise model” (NCNM) inspired by ordered categorical noise models. The noise model reflects an assumption that the data density is lower between the class-conditional densities. We illustrate our approach on a toy problem and present comparative results for the semi-supervised classification of handwritten digits.
TY  - CPAPER
TI  - Semi-supervised Learning via Gaussian Processes
AU  - Neil D. Lawrence
AU  - Michael I. Jordan
BT  - Advances in Neural Information Processing Systems
PY  - 2005/01/01
DA  - 2005/01/01
ED  - Lawrence Saul
ED  - Yair Weiss
ED  - Léon Bouttou	
ID  - lawrence-semisuper04
PB  - MIT Press	
SP  - 753
EP  - 760
UR  - http://inverseprobability.com/publications/lawrence-semisuper04.html
AB  - We present a probabilistic approach to learning a Gaussian Process classifier in the presence of unlabeled data. Our approach involves a “null category noise model” (NCNM) inspired by ordered categorical noise models. The noise model reflects an assumption that the data density is lower between the class-conditional densities. We illustrate our approach on a toy problem and present comparative results for the semi-supervised classification of handwritten digits.
ER  -

Lawrence, N.D. & Jordan, M.I.. (2005). Semi-supervised Learning via Gaussian Processes. Advances in Neural Information Processing Systems 17:753-760