Peter Orbanz


I am Assistant Professor in the Department of Statistics at Columbia University. Before coming to New York, I was a Research Fellow in the Machine Learning Group of Zoubin Ghahramani at the University of Cambridge, and previously a graduate student of Joachim M. Buhmann at ETH Zurich.

Tutorials on Bayesian nonparametrics

I have given a number of tutorials at NIPS 2011 (with Yee Whye Teh) and at Machine Learning Summer Schools.

Please see my tutorial page for slides, video recordings and further reading.


Stat W4400 (Statistical Machine Learning)


My main research interest are the statistics of discrete objects and structures: permutations, graphs, partitions, binary sequences. Most of my recent work concerns representation problems and latent variable algorithms in Bayesian nonparametrics. More generally, I am interested in all mathematical aspects of machine learning and artifical intelligence.

Working Papers

Lecture Notes on Bayesian Nonparametrics.
P Orbanz.
[PDF (draft)]

Bayesian Models of Graphs, Arrays and Other Exchangeable Random Structures.
P Orbanz and DM Roy.
[PDF (arXiv)]
Some of the key ideas in this survey are also explained in the following talk:

Borel liftings of graph limits.
P Orbanz and B Szegedy.
[PDF (arXiv)]

Unit-rate Poisson representations of completely random measures.
P Orbanz and S Williamson.

Nonparametric priors on complete separable metric spaces.
P Orbanz.


Random function priors for exchangeable arrays with applications to graphs and relational data.
JR Lloyd, P Orbanz, Z Ghahramani and DM Roy.
NIPS 2012.
[PDF] [Talk on this and related work (Video, 50min)]

Projective Limit Random Probabilities on Polish Spaces.
P Orbanz.
Electronic Journal of Statistics, Vol. 5, 1354-1373, 2011.

Dependent Indian Buffet Processes.
S Williamson, P Orbanz and Z Ghahramani.
AISTATS 2010, JMRL W&CP 9:924-931.

Bayesian Nonparametric Models.
P Orbanz and YW Teh.
In Encyclopedia of Machine Learning. Springer, 2010.

Construction of Nonparametric Bayesian Models from Parametric Bayes Equations.
P Orbanz.
NIPS 2009.
[PDF] [Supplements (Proofs)]
[Techreport Version] (Identical text; proofs included as appendix)

Music Preference Learning with Partial Information.
Y Moh, P Orbanz and JM Buhmann.
ICASSP 2008.

Nonparametric Bayesian Image Segmentation.
P Orbanz and JM Buhmann.
International Journal of Computer Vision (IJCV), Vol. 77, 25-45, 2008.
[PDF] [Publisher] [Code]

Cluster Analysis of Heterogeneuos Rank Data.
LM Busse, P Orbanz and JM Buhmann.
International Conference on Machine Learning (ICML), 2007.
[PDF (with corrections)]     [PDF (as published)]

Bayesian Order-Adaptive Clustering for Video Segmentation.
P Orbanz, S Braendle and JM Buhmann.
EMMCVPR, 2007.
[PDF] [Publisher]

Smooth Image Segmentation by Nonparametric Bayesian Inference.
P Orbanz and JM Buhmann.
European Conference on Computer Vision (ECCV), Vol. 1, 444-457, 2006.

SAR Images as Mixtures of Gaussian Mixtures.
P Orbanz and JM Buhmann.
IEEE International Conference on Image Processing (ICIP), Vol. 2, 209-212, 2005.
[PDF] [Publisher]

Notes and Techreports

Conjugate Projective Limits.
P Orbanz. [PDF (arXiv)]

Functional Conjugacy in Parametric Bayesian Models.
P Orbanz, 2009. [PDF]

PhD Thesis

Infinite-Dimensional Exponential Families in the Cluster Analysis of Structured Data.
ETH Zurich, 2008.


Nonparametric Bayes tutorials (various talks).
[Tutorial page]

Some other recent talks:

Nonparametric priors for exchangeable graphs and arrays.
[Slides] [Video]

Projective limit techniques in Bayesian nonparametrics.

Exchangeability, symmetry, and sufficiency.

Partition priors in computer vision.

Machine Learning Summer School 2009

While I was at Cambridge, we organized the Machine Learning Summer School 2009. All talks are available on Videolectures.

Iain Murray's talks on MCMC were my personal favorites on the program.


Room 1031 SSW
Department of Statistics, Columbia University
1255 Amsterdam Avenue
New York, NY-10027
Phone: +1-212-851-2148