Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!news.mathworks.com!hookup!lamarck.sura.net!howland.reston.ans.net!EU.net!sun4nl!aie.nl!news
From: geert@sparc.aie.nl (Geert-Jan van Opdorp)
Subject: Re: Artificial Forgetfulness
Sender: news@aie.nl (News Account)
Organization: AI Engineering BV
Lines: 40
Message-ID: <GEERT.95Apr19103232@sparc.aie.nl>
References: <3mj43i$300@ucthpx.uct.ac.za> <3ml747$qbi@news.xs4all.nl>
	<3musin$o0m@inferno.mpx.com.au> <3n0g5d$ce4@blackbird.afit.af.mil>
In-Reply-To: fkilpatr@afit.af.mil's message of Tue, 18 Apr 1995 13:46:46 GMT
Date: Wed, 19 Apr 1995 08:32:32 GMT


Hi,

>>>>> "Alex" == Alex Kilpatrick <fkilpatr@afit.af.mil> writes:

Alex> The interesting aspect of the game is the developers had to
Alex> implement some sort of Artificial Stupidity' to make the program
Alex> appear as slow as a human, while still being challenging.  The
Alex> Wodzap program always appeared to be just on the verge of
Alex> beating me, right before I won.

I once down-loaded a `memory' game (you know, with the pairs of
cards you have to locate) for my children. When the computer
played, it would turn the cards in a dull order, but moreover
it would never forget anything. This made it absolutely no
fun to play against. Ever since I've been wanting to 
implement a version that does forget. It should get confused
because of `similar' spatial relations of cards, visual
similarity (color, shape), and because of semantic
similarity (apple, pear). I did not give it much
thought yet, but maybe something like a combination
of semantic nets and `overloaded' associative networks 
would be useful.

Anyway, I agree it is fun to apply AI techniques in
such a `perverted' way. And useful for a Turing
test if our programs get too smart!

Geert-Jan






-- 
Geert-Jan van Opdorp
AI-Engineering
Amsterdam, The Netherlands
geert@aie.nl
