Name |
GRIMBERGEN, Reijer
Ph.D. (Cognitive Science)
Associate Professor |
Affiliation |
Department of Informatics
Faculty of Engineering
Yamagata University |
Specialization |
Cognitive Science
Human problem solving
Game programming |
Class Outline |
Games have been an important test environment in Artificial Intelligence
since the 1950s. The most notable success has been the 1997 victory of
IBM's "Deep Blue" chess computer against World Champion Gary
Kasparov. The next battle between human and computers will most likely
take place in the Asian games xiang qi (Chinese chess), shogi (Japanese
chess) or Go. At Yamagata University, we are doing research using the game
of shogi. It is expected that a shogi computer will beat the strongest
professional shogi player in less than ten years. In these lectures, first
the basic methods used in game programs will be explained. Then, a comparison
between chess and its Asian cousins will be made, followed by an explanation
of the new methods needed to challenge the top human players in shogi. |
Message to Students |
The techniques used in game programming are not limited to this domain
and neither are the techniques used in shogi programming limited to shogi.
This course will show what is needed to challenge human expertise with
computer models, using a very competitive domain (game programming is also
dubbed "the Formula 1 of Artificial Intelligence"). Also, the
explanations will be general enough such that knowledge of the rules of
the games used is not necessary. People with an interest in games, an interest
in the competition between man and machine or an interest in modeling human
thinking with computers are invited to follow this course. |
Theme |
1.Game programming methods (1): Minimax search and alpha-beta pruning
2.Game programming methods (2): Search optimizations
3.Computational comparison between different games
4.Special techniques for shogi (1): Mate search
5.Special techniques for shogi (2): Realization probability search and
the Yamagata University challenge
|