home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!cam-eng!ajr
- From: ajr@eng.cam.ac.uk (Tony Robinson)
- Newsgroups: comp.ai.neural-nets
- Subject: Re: NNs in chess or other games
- Message-ID: <1992Sep15.220149.6819@eng.cam.ac.uk>
- Date: 15 Sep 92 22:01:49 GMT
- References: <18psroINNnek@nestroy.wu-wien.ac.at>
- Sender: ajr@eng.cam.ac.uk (Tony Robinson)
- Organization: Cambridge University Engineering Department, UK
- Lines: 30
- Nntp-Posting-Host: dsl.eng.cam.ac.uk
-
- In comp.ai.neural-nets bruhn@uxe.wu-wien.ac.at (Peter Bruhn) writes:
-
- >Does anyone of you know something about an application of neural nets
- >in strategic games (esp. chess). What I think about is that a NN could
- >be trained to evaluate a position or some aspects of a position (e.g.
- >the structure of your pawns, how safe is the position of your king,...).
- >
- >.... Once a chess master (I don't remember his
- >name) was asked how many moves he prefigures in advance and he answered
- >"normally: none".
-
- There has been some interesting work done in this area. Perhaps the best
- complete system is for backgammon and is described in:
-
- Gerald Tesauro, "Practical Issues in Temporal Difference Learning",
- Machine Learning, vol 8, no 3/4, May 1992.
-
- The basic idea is to learn a connectionist position evaluator though self
- play. Supporting this are ideas from control theory, especially the backing
- up of lookahead using dynamic programming. A good reference is:
-
- Andrew G. Barto and Steven J. Bradtke and Satinder P. Singh,
- "Real-Time Learning and Control using Asynchronous Dynamic Programming",
- Department of Computer Science, University of Massachusetts, Amhurst MA,
- no 91-57, Aug 1991.
-
- Also, vol 8 of Machine Learning 1992 is a special issue on reinforcement
- learning and would be a good place to start.
-
- Tony [Robinson]
-