A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction: Insights and New Models

[edit]

Neil D. Lawrence, University of Sheffield

Journal of Machine Learning Research 13

Related Material

Abstract

We introduce a new perspective on spectral dimensionality reduction which views these methods as Gaussian Markov random fields (GRFs). Our unifying perspective is based on the maximum entropy principle which is in turn inspired by maximum variance unfolding. The resulting model, which we call maximum entropy unfolding (MEU) is a nonlinear generalization of principal component analysis. We relate the model to Laplacian eigenmaps and isomap. We show that parameter fitting in the locally linear embedding (LLE) is approximate maximum likelihood MEU. We introduce a variant of LLE that performs maximum likelihood exactly: Acyclic LLE (ALLE). We show that MEU and ALLE are competitive with the leading spectral approaches on a robot navigation visualization and a human motion capture data set. Finally the maximum likelihood perspective allows us to introduce a new approach to dimensionality reduction based on L1 regularization of the Gaussian random field via the graphical lasso.


@Article{lawrence-unifying12,
  title = 	 {A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction: Insights and New Models},
  journal =  	 {Journal of Machine Learning Research},
  author = 	 {Neil D. Lawrence},
  year = 	 {2012},
  volume = 	 {13},
  month = 	 {00},
  edit = 	 {https://github.com/lawrennd//publications/edit/gh-pages/_posts/2012-01-01-lawrence-unifying12.md},
  url =  	 {http://inverseprobability.com/publications/lawrence-unifying12.html},
  abstract = 	 {We introduce a new perspective on spectral dimensionality reduction which views these methods as Gaussian Markov random fields (GRFs). Our unifying perspective is based on the maximum entropy principle which is in turn inspired by maximum variance unfolding. The resulting model, which we call maximum entropy unfolding (MEU) is a nonlinear generalization of principal component analysis. We relate the model to Laplacian eigenmaps and isomap. We show that parameter fitting in the locally linear embedding (LLE) is approximate maximum likelihood MEU. We introduce a variant of LLE that performs maximum likelihood exactly: Acyclic LLE (ALLE). We show that MEU and ALLE are competitive with the leading spectral approaches on a robot navigation visualization and a human motion capture data set. Finally the maximum likelihood perspective allows us to introduce a new approach to dimensionality reduction based on L1 regularization of the Gaussian random field via the graphical lasso.},
  key = 	 {Lawrence-unifying12},
  linkpdf = 	 {ftp://ftp.dcs.shef.ac.uk/home/neil/spectral.pdf},
  linksoftware = {https://github.com/SheffieldML/meu},
  OPTgroup = 	 {}
 

}
%T A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction: Insights and New Models
%A Neil D. Lawrence
%B 
%C Journal of Machine Learning Research
%D 
%F lawrence-unifying12
%J Journal of Machine Learning Research	
%P --
%R 
%U http://inverseprobability.com/publications/lawrence-unifying12.html
%V 13
%X We introduce a new perspective on spectral dimensionality reduction which views these methods as Gaussian Markov random fields (GRFs). Our unifying perspective is based on the maximum entropy principle which is in turn inspired by maximum variance unfolding. The resulting model, which we call maximum entropy unfolding (MEU) is a nonlinear generalization of principal component analysis. We relate the model to Laplacian eigenmaps and isomap. We show that parameter fitting in the locally linear embedding (LLE) is approximate maximum likelihood MEU. We introduce a variant of LLE that performs maximum likelihood exactly: Acyclic LLE (ALLE). We show that MEU and ALLE are competitive with the leading spectral approaches on a robot navigation visualization and a human motion capture data set. Finally the maximum likelihood perspective allows us to introduce a new approach to dimensionality reduction based on L1 regularization of the Gaussian random field via the graphical lasso.
TY  - CPAPER
TI  - A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction: Insights and New Models
AU  - Neil D. Lawrence
PY  - 2012/01/01
DA  - 2012/01/01	
ID  - lawrence-unifying12	
SP  - 
EP  - 
L1  - ftp://ftp.dcs.shef.ac.uk/home/neil/spectral.pdf
UR  - http://inverseprobability.com/publications/lawrence-unifying12.html
AB  - We introduce a new perspective on spectral dimensionality reduction which views these methods as Gaussian Markov random fields (GRFs). Our unifying perspective is based on the maximum entropy principle which is in turn inspired by maximum variance unfolding. The resulting model, which we call maximum entropy unfolding (MEU) is a nonlinear generalization of principal component analysis. We relate the model to Laplacian eigenmaps and isomap. We show that parameter fitting in the locally linear embedding (LLE) is approximate maximum likelihood MEU. We introduce a variant of LLE that performs maximum likelihood exactly: Acyclic LLE (ALLE). We show that MEU and ALLE are competitive with the leading spectral approaches on a robot navigation visualization and a human motion capture data set. Finally the maximum likelihood perspective allows us to introduce a new approach to dimensionality reduction based on L1 regularization of the Gaussian random field via the graphical lasso.
ER  -

Lawrence, N.D.. (2012). A Unifying Probabilistic Perspective for Spectral Dimensionality Reduction: Insights and New Models. Journal of Machine Learning Research 13:-