Active Learning, Database Sampling, Experimental Design:
                     Views on Instance Selection

     Workshop held at the European Conference on Machine Learning
(ECML-2001) and the Conference on Principles and Practice of Knowledge
                Discovery and Data Mining (PKDD-2001).

          http://kd.cs.uni-magdeburg.de/~scheffer/ws01.html

                 September 4, 2001; Freiburg, Germany

                           Call For Papers
                           ===============

Workshop Scope
--------------

The  workshop   aims  at   bringing  together  researchers   from  the
communities of machine learning,  data mining, and statistics who are,
each with their own perspective, addressing a common problem which can
be referred to as instance selection.

In machine learning, researchers in the field of active learning study
algorithms that  actively control either  the amount of data,  or even
request data from  particular regions of the instance  space.  In Data
Mining,  the  term  sampling  is  used  to  describe  the  process  of
selecting, from  a database that is  too large to be  processed in its
entirety, a suitable  sample, ideally in such a way  that allows us to
give the  user guarantees  about how far  the obtained  results differ
from the optimal (non-sampling based) results.

In  statistics, finally,  we find  the foundation  for the  results of
active   learning  and  database   sampling,  starting   from  general
probability  tail  bounds  and  including  sequential  analysis  which
provides  the  basis  for  sequential sampling  algorithms.   Work  in
experimental design addresses the question  of how a minimal number of
data points can be selected that carries enough information to build a
good model of the underlying system.


Paper Submissions
-----------------

The program committee invites  submissions of original research papers
and extended abstracts as well  as papers describing work in progress. 
Contributions  are welcome  that relate  to active  learning, sampling
large  databases, or  related statistical  areas such  as experimental
design. 

Submissions should  be between two  and six pages  long.    Electronic
submissions are encouraged and should  be in postscript or PDF format. 
Please send  the paper, or  a URL pointing  at the paper, by  email to
scheffer@iws.cs.uni-magdeburg.de.  

Submissions must be received by 08 June 2001.


Important Dates
---------------

   08 June 2001:  Deadline for paper submission.
   29 June 2001:  Notification of acceptance.
   13 July 2001:  Camera-ready paper due.
   03 - 05 September 2001: PKDD-2001.
   05 - 07 September 2001: ECML-2001.
   04 September 2001: Workshop.


Organization
------------

Workshop Chairs: 
Tobias Scheffer and Stefan Wrobel, University of Magdeburg

Program Committee:
David Cohn, Carnegie Mellon University
Carlos Domingo, Synera Systems
Zoubin Ghahramani, University College London 
Hans-Joachim Lenz, Freie Universitaet Berlin
Gerd Paaß, GMD
Osamu Watanabe, Tokyo Institute of Technology 
Claus Weihs, University of Dortmund
Mohammed J. Zaki, University of Rochester