From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!spool.mu.edu!mips!zaphod.mps.ohio-state.edu!cis.ohio- Tue Mar 24 09:54:32 EST 1992
Article 4366 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!utgpu!csd.unb.ca!morgan.ucs.mun.ca!nstn.ns.ca!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!spool.mu.edu!mips!zaphod.mps.ohio-state.edu!cis.ohio-
state.edu!rutgers!rochester!cantaloupe.srv.cs.cmu.edu!crabapple.srv.cs.cmu.edu!andrew.cmu.edu!fb0m+
>From: fb0m+@andrew.cmu.edu (Franklin Boyle)
Newsgroups: comp.ai.philosophy
Subject: Re: Definition of Understanding
Message-ID: <Mdizvrm00WBME7ja4i@andrew.cmu.edu>
Date: 10 Mar 92 00:12:39 GMT
Organization: Cntr for Design of Educational Computing, Carnegie Mellon, Pittsburgh, PA
Lines: 30

Neil Rickert writes:

> The problem here is with the term "pattern matching".  This term implies,
>or at least suggests, that a current pattern is compared with a library
>of patterns for a match.  However things don't have to work that way.
>It is very likely that the brain is a "pattern recognition" system, but
>is not a "pattern matching" system.  That is, it uses various clues to
>recognize patterns, but there is no such thing as a library of stored
>patterns available for comparison.

I'm not sure what you mean by "various clues".  If you mean parts of
the pattern (i.e. partial matching), that is still pattern matching, it's
just that the tolerance for what constitutes a match is greater.
Also, one of my earlier points was that pattern recognition does not 
necessarily mean the physical process of structure fitting -- in digital
computers, however, it does.
 
> Likewise there is no reason computers can not be used for "pattern
>recognition" systems.  The fact that most recognition system are based
>on pattern matching is only a reflection on the limitations of current
>software.  There is no inherent reason that recognition must be based
>on matching.  Indeed there are many reasons to suspect that a good
>rapid recognition system with some ability to learn is the key to
>emulation of the behavior of the mind.

In digital computers, how do you recognize patterns if you physically don't 
couple one structure to another?  How is the pattern causal in the system 
*as that pattern*?

-Frank


