[robotics-worldwide] Call for Abstracts: RSS/PASCAL2 Workshop on Regression in Robotics

Christian Plagemann plagemann at stanford.edu
Wed Apr 22 16:44:20 PDT 2009


Workshop on Regression in Robotics -- Approaches and Applications

June 28, 2009, Seattle, WA, USA
Co-located with Robotics: Science & Systems
Sponsored by PASCAL2 network of excellence



May 1, 2009: Submission of poster abstracts
May 12, 2009: Notification of acceptances
June 28, 2009: Workshop


Please send an extended abstract (1 or 2 pages incl. figures) for a
poster presentation to submission at robreg.org

Partial travel funding for students is available through the generous
support of PASCAL2. Please indicate your interest in the submission

As a core PASCAL 2 event, the workshop will feature videotaped proceedings.

Authors of selected poster presentations will be invited to submit a
full length paper for an edited book to be potentially published by
Cambridge University Press.


Function approximation from noisy data is a central task in robot
learning. Relevant problems include sensor modeling, manipulation,
control, and many others. A large number of regression methods have
been proposed from statistics, machine learning and control system
theory to address robotics-related issues such as online updates,
active sampling, high dimensionality, non-homogeneous noise and
missing features. However, with minimal communication and
collaboration between communities, work is sometimes reproduced or
re-discovered, making research progress challenging.

Our goal is to draw researchers from the different communities of
robotics, control systems theory and machine learning into a
discussion of the relevant problems in function approximation to be
learned in robotics. We would like to develop a common understanding
of the benefits and drawbacks of different regression approaches and
to derive practical guidelines for selecting a suitable approach to a
given problem. In addition, we would like to discuss two key points of
criticism in current robot learning research. First, data-driven
machine learning methods do, in fact, not necessarily outperform
models designed by human experts and we would like to explore what
regression problems in robotics really have to be learned. Second,
regression methods are typically evaluated using different metrics and
data sets, making standardized comparisons challenging.

Goal & Topics:

We invite abstract submissions from researchers working on machine
learning, robotics and/or control theory with a general interest in
regression and function approximation. Ideally, submissions should
contribute to one or several of the following topics:

*** Approaches: Which learning approaches have been applied
successfully to solve regression problems in robotics or have a high
potential for doing so?

*** Problem settings: Which robot learning problems contain regression
or function approximation as a central component? What are the
specific aspects that make the problem challenging?

*** Theoretical foundations: How can challenging requirements such as
online updates, active sampling, high dimensionality, non-homogeneous
noise and missing features be addressed?

*** Benchmarking and evaluation: What are suitable methods for
evaluation of regression methods? What metrics are being used and,
subsequently, which should be used? Which benchmark data sets are
available and which are missing?

Workshop Organizers:

Christian Plagemann
Stanford University
plagemann at stanford.edu

Jo-Anne Ting
University of Edinburgh
jting at ed.ac.uk

Sethu Vijayakumar
University of Edinburgh
sethu.vijayakumar at ed.ac.uk

More information about the robotics-worldwide mailing list