Peter Orbanz


Welcome.

I'm a Research Associate (a postdoc) in the Machine Learning Group of Zoubin Ghahramani at the University of Cambridge. I hold a degree in computer science and mathematics from the University of Bonn, and obtained my PhD in the Pattern Analysis and Machine Learning Group at ETH Zurich, where I worked with Joachim M. Buhmann.


Contact

Peter Orbanz
Computational and Biological Learning Laboratory [link]
University of Cambridge

Email: p.orbanz@eng.cam.ac.uk
Phone: +44-1223-748518

Postal Address
Dr Peter Orbanz
University of Cambridge
Room BE441, Baker Building
Trumpington Street
Cambridge CB2 1PZ, UK


Research

I am interested in all mathematical aspects of pattern recognition and artifical intelligence. My current work focusses on the structural properties of Bayesian models, both in the parametric and nonparametric case.

Working Papers

        Conjugate Projective Limits.
        P Orbanz. Preprint.

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

        Construction of Nonparametric Bayesian Models from Parametric Bayes Equations.
        P Orbanz, 2009. Preprint. [PDF]

Publications

        Bayesian Nonparametric Models.
        P Orbanz and YW Teh. In Encyclopedia of Machine Learning (Springer), to appear.
        [PDF]

        Construction of Nonparametric Bayesian Models from Parametric Bayes Equations.
        P Orbanz. NIPS 2009.
        [PDF]         [Supplements (Proofs)]

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

        Nonparametric Bayesian Image Segmentation.
        P Orbanz and JM Buhmann. In 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. In 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. In EMMCVPR, 2007.
        [PDF]     [Publisher]

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

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

PhD Thesis

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


Tutorials, Teaching, etc

I have no teaching obligations at Cambridge, but I try to make up for it with tutorials and other ways to force myself upon the general public. I also keep some older teaching material available online.

Machine Learning Summer School 2009

We are organizing this year's edition of the Machine Learning Summer School here in Cambridge. [Website.]

Nonparametric Bayes Tutorial

A series of three tutorial talks on the theoretical foundations of nonparametric Bayesian methods. My tutorial webpage contains the talk slides, as well as an annotated list of references on Bayesian nonparametrics and related topics.

Teaching Material: Machine Learning

The teaching material (slides and exercise sheets) I drew up as a teaching assistant at ETH Zurich is available here:

Machine Learning Slides & Exercises

Judging from the surprising number of downloads, my exercise problems seem to be more popular than some of my papers.

Student's Theses

Theses I supervised at ETH Zurich.

        Patrick Pletscher: Model Order Selection: Criteria, Inference Strategies, and an Application to Biclustering.
        Master Thesis, ETH Zurich, 2007. [PDF]

        Ludwig Busse: Clustering Rank Data.
        Research Project, ETH Zurich, 2007.

        Samuel Braendle: Feature Extraction for Bayesian Order-Adaptive Clustering.
        Term Thesis, ETH Zurich, 2007. [PDF]

        Sarah Gugl: Classification of Radar Data Based on Constrained Agglomerative Segmentation.
        Term Thesis, ETH Zurich, 2006.