Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.sprintlink.net!EU.net!CERN.ch!news
From: Owen.RANSEN@cern.ch (Owen RANSEN)
Subject: Re: neural nets for backgammon
X-Nntp-Posting-Host: misf120.cern.ch
Message-ID: <D5qoK5.Dx3@news.cern.ch>
Sender: news@news.cern.ch (USENET News System)
Organization: CERN - European Organization for Nuclear Research
X-Newsreader: WinVN 0.93.11
References: <3k6hnq$l1c@smart.informatik.uni-kiel.de> <3k7ufc$sl@castor.anu.edu.au>
Mime-Version: 1.0
Date: Mon, 20 Mar 1995 12:40:53 GMT
Lines: 40

In article <3k7ufc$sl@castor.anu.edu.au>, lex@cs.anu.edu.au says...
>
>In article l1c@smart.informatik.uni-kiel.de, 
sp@informatik.uni-kiel.d400.de (Stefan Pantke) writes:
>>In <D5Foqn.E7C@news.cern.ch> ransen@cerncs2.cern.ch (owen ransen) 
writes:
>>
>>
>>>I read in Dr.Dobbs about a year ago about a neural network
>>>program which learned on it's own how to play backgammon.
>>>Does anywone know who wrote the NN version metioned in Dr.Dobbs?
>>>And are the sources freely available?
>>
>>Some years ago I read an article about a backgammon 
>>programm that played better than the world champion.
>>Look for the article in Scientific American.
>>
>>stefan
>
>
>I think the NN network system you are talking about is the one by 
Tesauro. It played
>against itself and adjusted the weights in the network to learn from this 
experience.
>Apparently the results achieved were quite good. I know the technical 
reports
>are available from WWW (from Computer Science archives), but am unaware 
of the
>code's availability.
>
>
>Lex
>
>
>
Thanks for your reply, and forgive my ignorange of the WWW, but what is
the URL address of the Computer Science archives, or even better the
actual page I am after?
Thanx

