Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!unmvax!ncar!tank!mimsy!eneevax!haven!aplcen!jhunix!ins_atge
From: ins_atge@jhunix.HCF.JHU.EDU (Thomas G Edwards)
Newsgroups: comp.ai
Subject: Re: Question on Chinese Room Argument
Summary: Sensory Basis of AI
Keywords: sensory
Message-ID: <855@jhunix.HCF.JHU.EDU>
Date: 24 Feb 89 03:34:03 GMT
References: <4298@pt.cs.cmu.edu> <Feb.18.17.26.17.1989.23438@elbereth.rutgers.edu> <4296@cs.Buffalo.EDU> <Feb.20.20.43.21.1989.15687@elbereth.rutgers.edu> <1989Feb20.213329.10376@cs.rochester.edu>
Reply-To: ins_atge@jhunix.UUCP (Thomas G Edwards)
Organization: The Johns Hopkins University - HCF
Lines: 16

During the Chinese Room discussion, many have brought up the
concept that sensory experience is the bottom rung of intelligence,
i.e. cat->chased by dog next door->dog's name is Fred->first letter
is "F"->"F" looks like the following sensory experience

While we may question the validity of the above idea, I'd like to
point out that an AI system does not need "real" sensory input in 
the sense of eyes, ears, etc., but can use internal enviromental
models (i.e. block's world).  This knowledge, though, needs to
somehow be entered into the computer from outside, in the form
of verbose description, or an algorithm (perhaps involving
pseudo-random numbers) describing the enviroment.

-Thomas Edwards

NN's on CM-2's!