Newsgroups: comp.robotics
Path: brunix!sgiblab!sdd.hp.com!elroy.jpl.nasa.gov!ames!agate!howland.reston.ans.net!newsserver.jvnc.net!stevens-tech.edu!vaxc.stevens-tech.edu!jcarnice
From: jcarnice@vaxc.stevens-tech.edu
Subject: <None>
Message-ID: <1994Mar27.023510.1@vaxc.stevens-tech.edu>
Lines: 26
Sender: news@dmi.stevens-tech.edu (USENET News System)
Organization: Stevens Institute Of Technology
Date: Sun, 27 Mar 1994 07:35:10 GMT

					<>

	  Because I believe that research into robotics should focus more on
	computers that learn and create knowledge representations for
	themselves than on researchers defining maps and specific frames of
	reference, I have been looking at the artificial life field.	

	  I don't see Von Neuman machines being effective in promoting
	"learning" using the software philosophy commonly used.  I think
	creating virtual worlds with self-evolving entities might shed some
	light on the field of artificial intelligence and robot control.

	  Nothing so extravagant as "let's evolve these things until they're
	totally sentient and then put them into robot bodies."  It's easier
	to wonder if, instead of mutating and selecting individual based upon
	simple requirements, one might not mutate new "ideas" and select
	them based upon simple tests of consistency with other ideas.

	  Could an intelligent computer consist of "living" and "evolving"
	ideas in such a way that's stable and self-organizing?  That's
	what I'm studying and want to know.  Anyone else doing similar
	research?

						-James Vincent Carnicelli II
						 Stevens Institute of Technology
						 Hoboken, New Jersey
