I am very interested in computer chess. I'm a mere first year
undergraduate now, but I'm planning as far ahead as my Ph.D., where I'd
like to concentrate on computer chess.
The thing that interests me most is the knowledge-based approach as
opposed to the DT-type numbercrunching (tell me one Grandmaster that
goes through 1.6 million positions in one second). Although it might not
be as successful now, it is certainly more interesting and a lot more
related to Artificial Intelligence.
The most interesting paper I have come across so far is the following:
D.E.WILKINS: USING PATTERNS AND PLANS TO SOLVE PROBLEMS AND CONTROL
SEARCH, Stanford Uni research memo AIM-329/CS 79-747 (1979).
In it is described a program called PARADISE which uses a knowledge base
to solve tactically sharp positions. He suggests several methods of
improving it, e.g. speeding it up etc.
I haven't heard of anybody who's picked up on it, and nowadays the
interest seems to lie in special hardware/parallel processing ideas.
Could someone tell me what has happened since 1979?
Also, I wonder whether it would be possible to somehow get the
sourcecode for PARADISE (maybe through my Department?) or is that too
big a blasphemy?
Also, any hints as to what else would be interesting will be greatly
appreciated.
--
Nicolai Czempin has no fancy .signature yet.