content-transfer-encoding: quoted-printable
---------------------------------------------------------------------------=
-----
Hello Ian et. al.,
I mentioned an article from Scientific American (from on or about 1987) cal=
led "Collective Computation in Neuronlike Circuits" a couple weeks ago that=
addreses the subject of your post. It was written by David W. Tank and Jo=
hn J Hopfield. I've yet to have the opportunity to scan this so the rest o=
f the list could have it to read. SciAm dosen't even have a copy of it! =
The flip flop circuit in the article is similar to a BiCORE, so I made the =
logical assumption that it was a ProtoBEAM circuit. Attached is my renderi=
ng of the circuit. The values are non-essential, as they're the ones that =
I always start with when building BEAM circuits (default, if you will). I =
built this one with a 74HC14, but it didn't work very well. I've yet to tr=
y a '240. I've even given thought to using a '245. =
My thoughts on this are to use a '240 or some such thing like that to use a=
s a 'plug in' memory unit that can be chained together to form larger memor=
ies. Also these memories (ideally) would be externally as well as internal=
ly initiated. By this I mean that several BEAM robots would be IR linked t=
ogether (or RF, whichever is easier) in order to make up a larger secondary=
neural network or collective network. =
This kind of thing has been bandied about quite a lot on the list, but no a=
ctual circuitry has been offered. All that seems to go on is talk...Reg wi=
ll not be taking part in any terrorist action as he has a bad back...that s=
ort of thing. Let's work together and see what we can come up with. =
It is my sincerest hope that none will be bored by this line of discussion.=
=
=
Richard Caudle
www.geocities.com/frankendaddy
Home ICQ - Frankendaddy
---------------------------------------------------------------------------=
-----
Guardians of the sacred words: Nee, Ptang, and NeeWhon! =
content-transfer-encoding: quoted-printable
>
Hello Ian et. al.,
I mentioned an article from Scie=
ntific =
American (from on or about 1987) called "Collective Computation in Neuronli=
ke =
Circuits" a couple weeks ago that addreses the subject of your post.&n=
bsp; =
It was written by David W. Tank and John J Hopfield. I've yet to have=
the =
opportunity to scan this so the rest of the list could have it to read.&nbs=
p; =
SciAm dosen't even have a copy of it!
The flip flop circuit in the art=
icle =
is similar to a BiCORE, so I made the logical assumption that it was a=
=
ProtoBEAM circuit. Attached is my rendering of the circuit. The=
=
values are non-essential, as they're the ones that I always start with when=
=
building BEAM circuits (default, if you will). I built this one with =
a =
74HC14, but it didn't work very well. I've yet to try a '240. I=
've =
even given thought to using a '245.
My thoughts on this are to use a=
'240 or =
some such thing like that to use as a 'plug in' memory unit that can be cha=
ined =
together to form larger memories. Also these memories (ideally) would=
be =
externally as well as internally initiated. By this I mean t=
hat =
several BEAM robots would be IR linked together (or RF, whichever is easier=
) in =
order to make up a larger secondary neural network or collective =
network.
This kind of thing has been band=
ied about =
quite a lot on the list, but no actual circuitry has been offered. Al=
l =
that seems to go on is talk...Reg will no=
t be =
taking part in any terrorist action as he has a bad =
back...that sort of thing. Let's work together a=
nd =
see what we can come up with.
It is my sincerest hope that non=
e will be =
bored by this line of discussion.
Guardians of the sacred words: N=
ee, =
Ptang, and NeeWhon! =
Attachment: sciamflipflop.jpg
Home