[robotics-worldwide] [meetings] CFP: RSS 14 Workshop on Non-parametric Learning in Robotics

Rudolph Triebel rudolph.triebel at in.tum.de
Fri May 2 06:12:41 PDT 2014


======================================================================
                        CALL FOR PAPERS

     RSS 14: Workshop on Non-parametric Learning in Robotics

             Submission deadline:      June 6,  2014
             Notification:             June 30, 2014
             Workshop date:            July 12, 2014

       http://ais.informatik.uni-freiburg.de/nonparam_rss14

======================================================================

Abstract:

The  growing interest  in non-parametric  machine learning  methods is
driven by  their flexibility and expressive  power on one  side and by
their efficiency  when applied to large  data sets on  the other side.
The latter is particularly interesting for robotic learning tasks, and
recent achievements show the potential  that these methods can have in
practice. In  this workshop,  we will present  non-parametric learning
methods  including Gaussian  Processes,  Spectral Learning,  Dirichlet
Processes, Deep  Learning, and we will show  potential applications in
robotics. Renowed  experts in the  field will present their  work, and
there will be ample  opportunities for interaction and discussion. The
aims are to draw further  attention of the robotics community to these
novel  methods,  and  to   highlight  their  benefits  over  standard,
parametric learning techniques.

Invited Speakers (subject to change):

 - Michael Jordan, Univ. California, Berkely, USA
 - Ashutosh Saxena, Cornell Univ., USA
 - Kristian Kersting, TU Dortmund, Germany
 - Jonathan P. How, MIT, USA
 - Byron Boots, Univ. of Washington, USA

Submissions:

We solicit submissions in all areas related to non-parametric learning
methods  for  robotics   applications,  including  Gaussian  Processes
(regression  and classification),  Dirichlet  Processes and  variants,
Deep  learning, Support Vector  Machines, etc.  In particular,  we are
interested  in applications  to  important robotics  problems such  as
navigation, planning, mapping, manipulation and exploration.


Format:

Submissions can  be either extended abstracts  of previously published
work  relevant  to the  field,  or  novel  contributions. An  extended
abstract should not exceed 3  pages of standard RSS paper style, novel
contributions can have a length of  up to 6 pages. Please indicate the
type   of  submission   in  the   document.   All   contributions  are
peer-reviewed, and  the accepted papers are collected  in the workshop
proceedings, which will be available on the work shop web page. Please
upload your submission to the following site:

     https://www.easychair.org/conferences/?conf=nparmlearn14


Presentation:

Authors of all accepted papers are  invited to present their work in a
poster session during the work shop. In addition, a selected number of
contributions will be given the opportunity for oral presentations. In
concordance with  the speakers, the  talks will be video  recorded and
made available to the public after the work shop.

Organizers:

Rudolph Triebel, TU Munich, Germany (rudolph.triebel at in.tum.de)
Luciano Spinello, Univ. Freiburg, Germany (spinello at cs.uni-freiburg.de)





More information about the robotics-worldwide mailing list