From: Eugene Leitl (Eugene.Leitl@lrz.uni-muenchen.de)
Date: Sun Jan 05 1997 - 11:36:35 MST
The latest and most exciting development in ATR's CAM-Brain Project is
the very recent (September 96) brainstorming work with an American
based hardware designer on the so-called "CAM-Brain Machine (CBM)",
which should be implemented by summer of 97, to evolve neural net
modules in less than a second (with a genetic algorithm population of
100 and with 100 generations or so). If this project is approved by
ATR management, this section will be updated in greater detail. The
excitement comes from the realisation that a CBM will make brain
building (i.e. evolving a million neural net modules and putting them
together into artificial brains) practical and realistic. For the past
few years, the general reaction to the CAM-Brain Project has been one
of "wait and see, tinged with a bit of scepticism, due to the large
numbers of neurons mentioned (i.e. a billion)". CBMs will change all
this. Personally, I believe that CBMs will give birth to a new
industry - artificial brains, brain-like computers. I am very excited.
UPDATE 24 December 1996
A contract between ATR and Dr. Mikhail KORKIN was signed Xmas Eve,
1996. A prototype version of CBM will be delivered by April 1997, and
if it passes appropriate feasibility tests, the full machine should be
delivered to ATR by the end of 1997. This machine will be based on
Xilinx's XC6264 chips, and will evolve a neural net module in a
fraction of a second, thus making brain building practical.
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:43:57 MST