We need to put together a list of prospective attendees of this workshop. If you
might be interested in attending, please send me email.
Thanks,
Lewis Johnson
------- Forwarded Message
Proposal for IJCAI-95
Workshop on Agent Modeling
The study of issues arising in building agents competent in
interacting with other agents has played a central role in much of AI
research. The reason for this interest is the fact that an automated
intelligent agent often needs to interact with other agents in its
environment to achieve its goals. Agent modeling -- a capability to
model and reason about other agent's knowledge, beliefs, goals, and
actions -- is key in intelligent interaction. This capability is being
addressed in a variety of research areas, including distributed AI
(DAI) and multi-agent systems, natural language discourse, plan
recognition, intelligent tutoring and user interfaces, as well as in
some areas of theoretical economics, such as game theory. A variety
of techniques are being used, from building theoretical models to
implemented systems, within all of the areas. Added impetus to this
work comes from the recent explosion of work on agents in dynamic
interactive (virtual reality) simulation, software (such as
information retrieval), and related environments. These environments
bring with them a new set of concerns for agent modeling for
collaboration and competition.
Aims of the Workshop and Themes
The workshop is intended to bring together researchers working in all
of the areas related to the theoretical and practical aspects of agent
modeling. The aim is to assess the state of the art, discuss common
foundations of representation and reasoning with models of agents, and
explore future directions in which current theories and systems must
be extended. We hope that a productive (and provocative) exchange of
ideas will further the participants' understanding and expertise.
Specific topics include, but are not limited to:
\begin{itemize}
\item Are theoretical models developed for agent modeling in DAI of relevance
for building intelligent tutoring systems?
\item How do concerns of real-time performance and reactivity in the synthetic
simulation environments change the nature of agent modeling?
\item What level of modeling is appropriate for different tasks? Is
it always the case that treating agents as equipped with knowledge,
beliefs and desires is better? Can they be treated as simple
mechanisms?
\item What representations emerge as common across the subfields and
applications? Are there any promising alternatives?
\item What are the methods of making use of the models?
How is the information contained in the model being processed?
\item Can the models considered in game theory be used in AI
applications?
\end{itemize}
Note that while the AAAI-95 spring symposium on "Representing
mental states and mechanisms" addresses a somewhat related topic,
their focus is on representations of self-knowledge and of
sharability and usability of these representations rather than
on the task of agent modeling.
Suggestions for encouraging a workshop-like rather
than a mini-conference-like atmosphere:
A hindrance to workshop-like
atmosphere is often a schedule that includes lengthy talks
uninterrupted by periods of question/answers or discussions. We
propose to follow a format whereby attendees presenting their work
will be given a relatively short time period to cover salient points
in their work, and then an almost equal amount of time for a
question-answer period -- perhaps 10 minutes for talk plus 10 minutes
for questions. Invited speakers or members of the organizing committee
may also lead discussion sessions bringing up key research problems.
Depending on the number of submissions, a poster session may also be
organized. However, submission of a paper is not a pre-requisite for
participation in the workshop. Interested researchers will be welcome
subject to space limitations.
The deadline for submissions and requests
for participation is Feb 10, 1995. Participants will be notified
of acceptances by March 10, 1995. They have until March 31, 1995 to
submit a final version of the their paper. This gives us a week
week to submit to IJCAI by April 7, 1995: (a) a provisional list
of workshop participants and (b) working notes to be duplicated
for the workshop.
Organizing committee:
Milind Tambe (co-chair) Piotr Gmytrasiewicz (co-chair)
Information Sciences Institute Computer Science and Engineering
University of Southern California University of Texas at Arlington
4676 Admiralty Way Arlington, TX 76019
Marina del Rey, CA 90292 USA
USA piotr@cse.uta.edu
tambe@isi.edu (817) 273-3334
(310) 822-1511
Ed Durfee Lewis Johnson
Electrical Engineering and
Computer Science Information Sciences Institute
AI Lab University of Southern California
University of Michigan 4676 Admiralty Way
Ann Arbor, MI 48109 Marina del Rey, CA 90292
USA USA
durfee@eecs.umich.edu johnson@isi.edu
(313) 936-1563 (310) 822-1511
Charles Rich Candice Sidner
Jeffrey Rosenschein John Self
Computer Science Department
Hebrew University
Jerusalem, Israel
e-mail: jeff@cs.huji.ac.il
------- End of Forwarded Message