*** First Call for Papers ***

                     AAAI Fall Symposium 2000


                    LEARNING HOW TO DO THINGS

                        November 3-5, 2000
                     Cape Cod, Massachusetts


                    www.dfki.de/~bauer/fs2000

Symposium Topic

Knowing how to do things is an important  category  of  knowledge
underlying  many  kinds  of  intelligent  behavior  in artificial
agents,   such   as   critiquing,   advice   giving,    tutoring,
collaboration,  and  delegation. In the current state of the art,
most of this procedural knowledge  is  encoded  "manually"  by  a
single  person  (or  a small team) who needs to be expert in both
the task domain  and  the  appropriate  knowledge  representation
formalisms.  This  is  a serious bottleneck in the development of
these kinds of systems.

The focus of this symposium is on how to  automate  or  partially
automate the acquisition of procedural knowledge, namely, indexed
collections  of  what  are  variously   called   macros,   plans,
procedures,  or  recipes for action. The techniques for acquiring
this knowledge may depend on many variables, including:

   * size of the domain (e.g., number of recipes)
   * amount of input data
   * number of steps in a typical task
   * type of tasks (e.g., analysis vs. synthesis)
   * number of agents involved (e.g., one, two, or many)
   * type of agents involved (e.g., human vs. computer)
   * intended use of the  knowledge  (e.g.,  acting,  critiquing,
     etc.)
   * degree  of  supervision  (e.g.,  teaching  vs.  unsupervised
     learning)
   * level of abstraction (e.g., primitive operations  vs.  high-
     level goals)
   * degree of  initiative  (e.g.,  learning  by  experimentation
     versus passively)

Because  of  this  problem  diversity,   we   hope   to   include
participants  in  the  workshop  from a number of research areas,
including:

   *  programming  by  demonstration  (highly  supervised,  small
      amount of input data)
   *  data mining (unsupervised, large amount of input data)
   *  case-based  problem  solving  (cases  are   like   recipes,
      especially if abstracted)
   *  machine learning (range of techniques)
   *  cognitive  and  social  sciences  (e.g.,  studies  of  human
      instructional dialogues)
   *  instructable agents


Important Dates

           Submission of position papers:       March 29, 2000
           Notification of acceptance:          May 5, 2000
           Registration deadline:               May 25, 2000
           Submission of final paper versions:  August 25, 2000
           Symposium:                           November 3-5, 2000


Submission

Potential participants  should  submit  a  short  position  paper
(maximum three pages) containing the following four elements:

  1. Primary  contact:  name,  affiliation,   postal   and   email
     addresses, telephone and fax numbers. Invitations to  second-
     ary authors will be made only if they are also listed on this
     submission.
  2. Statement and discussion  of two or three important  research
     questions  that  could  be  presented  and  discussed  at the
     workshop.
  3. Statement and discussion of a  domain that could serve  as  a
     shared  example for the workshop. Explain how this particular
     domain  would  help  make  our  discussion  more concrete and
     productive.
  4. A  short  summary  of  authors'  relevant   work,   including
     references (please supply URLs if available).

Please   email   submissions   (plain   ascii   text   only)    to
learninghow@dfki.de.  Confirmation  of receipt will be returned by
email.

Organizing Committee

Mathias  Bauer,  DFKI  (bauer@dfki.de,  co-chair)
Charles  Rich, Mitsubishi Electric Research Lab. (rich@merl.com,
co-chair)
Andrew Garland, Brandeis University
Abigail  Gertner,  University Pittsburgh
Eric Horvitz, Microsoft Research
Tessa Lau, University Washington
Neal Lesh, Mitsubishi Electric Research Lab.
James Lester, North Carolina State University
Henry Lieberman, MIT
Jeff Rickel, USC/ISI
Candace Sidner, Lotus Development Corp.