Loading ...
Sorry, an error occurred while loading the content.
 

CFP: Learning with Nonparametric Bayesian Methods - ICML 2006 Workshop

Expand Messages
  • steffenbickel
    ******************************************************************** CALL FOR PAPERS / ABSTRACTS ICML 2006 Workshop Learning with Nonparametric Bayesian
    Message 1 of 1 , Mar 31, 2006
      ********************************************************************

      CALL FOR PAPERS / ABSTRACTS

      ICML 2006 Workshop

      Learning with Nonparametric Bayesian Methods

      Pittsburgh, Pennsylvania, June 29, 2006

      Deadline for submissions: April 28, 2006

      ********************************************************************

      INTRODUCTION

      Dirichlet Processes and other nonparametric Bayesian (NPB) methods
      have originally been developed in statistics but are finding
      growing interest in the machine learning community. Although the
      name indicates otherwise, NPB is concerned with models with an
      infinite number of parameters. For machine learning practitioners
      this leads to attractive models with (countably) infinite
      dimensions in a hidden state space like infinite mixture models.
      NPB models have the favorable property that their complexity
      automatically adapts to the number of data points. It has already
      been demonstrated that in some important machine learning
      applications, NPB has clear advantages over parametric solutions.
      We hope that this workshop will serve as a platform to discuss
      basic issues and recent developments in NPB.


      TOPICS AND QUESTIONS WE WANT TO ADDRESS

      * General principles:
      + We plan an introductory talk on nonparametric Bayesian methods.

      * Current developments:
      + What are the recent developments in the field of NPB?
      + Are there interesting new applications?

      * Open problems/new challenges:
      + What are the problem settings for which satisfactory NPB
      solutions are still missing due to modeling or inferential
      issues?
      + In which areas NPB methods could not demonstrate superior
      performance, if compared to parametric solutions?
      + Are there any new challenges arising from recent developments
      like spatial, time-varying or transformed Dirichlet processes?

      * Computational issues:
      + How can we improve the speed of parameter estimation and
      inference?
      + What is the right estimation/inference method for what
      setting (MCMC, variational Bayes, empirical Bayes, expectation
      propagation)?
      + Are we ready for large data sets, high dimensional data, or
      online data processing?


      PAPER/ABSTRACT SUBMISSION

      We strongly encourage researchers in the area of machine learning,
      statistics, natural language processing, computational biology,
      information retrieval, and related fields to either submit an
      extended abstract (less than 2000 words) or a full paper (4-8
      pages). Each submission will be reviewed by at least two reviewers.

      Please submit your abstract or paper electronically (PDF or
      postscript format) to bickel@...-berlin.de. It is
      recommended to submit papers using the ICML 2006 conference paper
      style. Submissions should include the names and contact information
      of the authors.


      WORKSHOP FORMAT

      This will be a one-day workshop immediately after the main ICML
      conference. The workshop will interleave invited talks and technical
      presentations of the accepted submissions with extensive time for
      discussion of the presented work.


      IMPORTANT DATES

      April 28, 2006: Abstract and paper submission deadline
      May 19, 2006: Notification of acceptance
      June 09, 2006: Camera ready copy deadline for online workshop
      proceedings
      June 29, 2006: Workshop


      ORGANIZING COMMITTEE

      Steffen Bickel
      Humboldt University, Berlin, Germany

      Volker Tresp
      Siemens AG, Corporate Technology, Munich, Germany


      PROGRAM COMMITTEE

      - Michael Jordan, University of California, Berkeley
      - Zoubin Ghahramani, University of Cambridge
      - Michael Escobar, University of Toronto
      - David Blei, Princeton University
      - Yee Whye Teh, National University of Singapore
      - Matthew Beal, State University New York, Buffalo
      - Thomas Griffiths, Brown University
      - David Draper, University of California, Santa Cruz
      - Athanasios Kottas, University of California, Santa Cruz
      - Larry Wasserman, Carnegie Mellon University
      - Kai Yu, Siemens AG, Corporate Technology
      - Wray Buntine, Helsinki Institute of Information Technology
      - Eric Xing, Carnegie Mellon University
      - Jerry Zhu, University of Wisconsin-Madison


      For more information, please visit
      http://www.informatik.hu-berlin.de/~bickel/npb_workshop.html

      We are looking forward to an interesting workshop and encourage
      your participation.

      Volker Tresp and Steffen Bickel
    Your message has been successfully submitted and would be delivered to recipients shortly.