---------------------- Forwarded Message: ---------------------
From: Hugo de Garis <degaris@starlab.net>
To: CT <tilley@worldnet.att.net>, degaris@starlab.net
Subject: Re:
Date: Mon, 07 Feb 2000 20:11:20 +0100
Dear C. Tilley,
Figuring out how to put 64000 evolved (1000 neurons each) circuit modules
together
is now my biggest challenge and will give me headaches in the year 2000 and
beyond.
Since the CAM-Brain Machine (CBM) had only just started working, I dont yet
have any
experience on putting many modules together to see what they can do. Once the
machine is
delivered to my new lab (STARLAB, http://www.starlab.org) May/June 2000, then
my team and
I can start the real brain building work. Several dozen "evolutionary engineers
(EEs)" will be needed to
specify the functions of all the modules, and several "brain architects (BAs)"
will be needed to
provide the overall design. Brain building will certainly NOT be random. What
happens within a
module is random however (provided the module produces a high fitness given the
criterion for evolution)
. I have no idea how my evolved circuits do their thing. In many ways I dont
care.
Each one is too complex and there will be too many of them. As we play with
multimodule systems,
we will get a feel for what can be done. I'll put progress reports on my web
sites as the results and lessons come in.
http://www.cs.usu.edu/~degaris
http://foobar.starlab.net/~degaris
Cheers,
Prof. Dr. Hugo de Garis
degaris@starlab.net
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:03:31 MDT