| |
This semester, DIST faculty member Paul Munro is leading
a doctoral seminar on "Interactive Agents in
Uncertain Environments", for which the development of
Poker-playing software is a primary focus. He points
out that, "A good Poker player must assess the other
players in order to interpret their actions and make
reasonable inferences about their hands. At the same
time, a good player should not behave in a way that
gives away their own hands. Thus the AI that goes
into a Poker player has many levels and is conceptually
more challenging
than a game like Chess."
The idea for the seminar was inspired by a symposium
held at the 6th International Conference on Cognitive
Modeling (ICCM), which was cohosted by Pitt and CMU
this past July. The theme of the symposium was to build
Poker-playing software agents. At the conference
banquet, held at Dave & Buster's, a Poker game between
humans and the software agents was held and displayed
for the audience a'la ESPN Poker on TV.
Dr. Munro was a member of the ICCM organizing committee,
along with Christian Schunn (LRDC), Marsa Lovett
(CMU Psychology), and Christian Lebiere (CMU HCI
Institute). The meeting included several oral presentations,
poster presentations, tutorials, and a doctoral consortium.
In addition to the so-called "Pokerbot" symposium,
highlights of the conference were a performance piece
dialog between a human and a "robot" and keynote talks
by Ken Forbus (Northwestern) and Michael Mozer (University
of Colorado). For more details see http://simon.lrdc.pitt.edu/~iccm/.
|
|
