Computer Chess Club Archives


Search

Terms

Messages

Subject: Learning and doing research

Author: Russell Reagan

Date: 22:56:13 03/19/03


I am interested to learn about alternatives to what has become the more or less
"standard" way that chess programs work (and other similar game playing
programs). I have recently written programs to play games other than chess, and
I have found that the "standard" methods used to write a good chess program
don't produce the same results in other games. The question that got me thinking
about this was, are we reaching the limits of what alpha-beta based algorithms
can do for us? I know of maybe a couple of strong chess programs that you might
not call "standard", but I think most programs use some form of alpha-beta,
along with some kind of pruning, various heuristics, and so on. A great amount
of research has been done on alpha-beta and related algorithms. What about other
alternative approaches? I wonder if they have been studied as thoroughly.

So how does one conduct research? I assume that the first task is to familiarize
yourself with what has been done in the past. So how does one go about doing
this? There are academic papers all over the net, and probably many more
scattered across university libraries. How is one supposed to know what is even
out there to read? If I went to my local city library, they probably aren't
going to have archives of artificial intelligence journals, and I'm sure that
the amount of archives varies even from university to university. What papers
have been written? Where can you find the papers? Do you have to hope the
professor (or whoever) that wrote the paper has a website with it available for
download? Or do you have to go hunting at numerous universities until you find
one that has what you need and get a copy from them? Then there are
organizations like ICGA that have a lot of good papers which you can get. It
seems like the information is very scattered about.

I know there are people here who have done research and written papers, and I
would like some guidance as to how to conduct research and learn about what has
been done and researched in the past, and how to get "up to speed" on a
particular domain of study.

Thanks,
Russell



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.