Computer Go
Table of contents |
Introduction
There is much yet to be done in the field of computer Go. While many different approaches have been tried, the level of the best Go playing programs is still quite low not exceeding amateur 1 dan level. An amateur dan player, who has learned the weakness of specific programs, can beat computer go at nine stones, not to mention professionals. [1]
Thus computer Go is a very exciting part of Artificial Intelligence (AI), many new ideas and techniques are yet to be discovered. It's interesting to note that compared to the approach which proved so successful with Chess, brute force, appears to be not nearly as useful for solving Go problems given the huge branching factor of the game tree.
In fact, Go endgames have been proven to be PSpace-hard, let alone other parts of the game. Also, many other aspects of Go, including life and death, are also known to be NP-hard. This means that it is very unlikely to be able to find a reasonably fast algorithm for playing perfect Go. So it looks like it's all about heuristics (surprise surprise).
Patrick Taylor: I'm not sure heuristics are all that bad. Humans don't know the perfect move sequence any more than computers do. Therefore, we basically play with heuristics as well. Proverbs are essentially heuristics used by human players to approximate good play.
References
See the Computer Go Bibliography maintained by the
Computer Go Group at the
University of Alberta.
Philosophy
Before delving into computer Go, it would behoove the novice to consider some philosophical questions about computers and Go.
List of Existing Go Playing Programs
Please see Go Playing Programs for a discussion about the best programs currently available.
List of Existing Problem Solving Programs
- Go Tools -- enclosed life and death problems
- MadLab -- tesuji problems
Robots
Some online Go servers, such as KGS, provide software opponents or robots to clients. Robot is derived from the Czech word robota meaning drudgery or slave labour which certainly describes the work needed to pummel double digit kyu players like me. PatG
Anti-Computer Strategy
It might be entertaining and educational to consider strategies for beating computers at Go.
Competitions
Below are competitions where Go playing programs can be tested.
International Competitions:
- Computer Olympiad (1988-present)
- Gifu Challenge (2003-2006)
- 21st Century Cup (2001-2002)
- Ing World Championship (1985-2000)
- Winners could make an attempt at the Ing Prize
- FOST (1995-1999)
Regional Competitions:
- European Championship (1987-2005,2008)
- US Championship (1988-2000,2008)
- USENIX (1984-1988)
Small Board Competitions:
-
2002 Computer-Go 13x13 Tournament -- Osnabrueck
- CGOS
- monthly KGS tournaments
- Computer Go Ladder
History of Go Programs
- [1989]: GNU Go 1.1 was posted to comp.sources.games March 13 1989
- [1968]: Albert L. Zobrist wrote the first ever program which played complete Go games.
- Ishi - old file format
Articles in Magazines
-
To Test a Powerful Computer, Play an Ancient Game
- Go Computings Future Article
-
Cracking Go by Feng-Hsiung Hsu published in the October 2007 issue of the IEEE magazine
Spectrum Online.
Links
-
http://www.computer-go.info/h-c/index.html contains a list with result of computer vs professionals.
-
http://www.citeulike.org/group/5884/library (Computer Go Bibliography)
(Smaller Go Bibliography, with comments)
-
http://gobase.org/go-7.html : Jan van der Steen's computer Go page has numerous links to famous programs and programmers as well as programming resources and articles.
-
http://www.epp.infonomics.nl/FdAW/chen/go.htm : A nice compilation of the strategies and heuristics used by current Go playing programs. (LINK BROKEN)
-
http://nis.acs.uci.edu/~strombrg/ml.html (Machine Learning) Including Neural Networks, Genetic Algorithms, Temporal Difference Learning. (LINK BROKEN)
-
http://www.cgl.ucsf.edu/go/ladder.html (The Computer Go Ladder) still active ?
Maillist(-Archives)
- gnugo-list
-
http://computer-go.org/pipermail/computer-go/ Active computer-go list archives.
-
http://www.daogo.org/ Archive of Computer Go newsletter, issues 1 to 16. In PDF.
Authors: Gounter, Chestnut
Notes:
[1] Will a Go Program Defeat the Meijin by Aratani Kazunari.
Bill: Note on the article. Typically, when computer go tournaments are held in Asia, some child will play a program an even or low handicap game, and the program does reasonably well. Claims about the strength of programs are made, based upon such results. Compare that with the claims about the strength of the tsumego solver, Go Tools, which has been thoroughly tested over the years.