FleshFactor FleshFactor
logo; AEC FORUM

[Prev][Next][Index][Thread]

FleshFactor: Re: to Reply to Some Scientific Angles



---------------------------------------------------------
A E C  F O R U M - "F L E S H F A C T O R"
(http://www.aec.at/fleshfactor/arch/)
---------------------------------------------------------


In response to the following paragraph which I wrote in "Some 
Scientifc Angles"...

>THEORY (and I mean VERY fundamental, from a physics perspective!):

>1) Some people say that the brain works using tiny neurons which work
>using chemistry and electrical impulses which both use physics. If we
>could understand exactly how this chemistry works, then we could, in
>theory, make something that works exactly like our brains. Furthermore,
>since these processes can, in theory, be modelled on a computer, we
>could use a vast network of big computers to model the chemistry of the
>neurons and of their interconnections so as to make a brain in a
>computer.

.....Mark Weiser replied with:

>This is similar to an argument that if I understood exactly how all the
>transistors in a computer worked, then I could make something that works
>exactly like the computer.  What is wrong with this argument?  Well,
>anyone heard of software?  Anyone think that if they know about all the
>transisters and their interconnections that they understand anything
>about the software on the computer?  A computer without software is
>indistinguishable from a hot rock, except in its potential. 


I think this is a misunderstanding of what I meant to say so perhaps I 
should clarify, since I don't always make sense to others when I think 
I am!  I agree with the rest of the comments in that letter but I 
would like to add another of my own -- in attempting to dig up a 
"fundamental" theory we will inevitably find that there is something 
even more fundamental underneath it which we cannot understand at the 
moment. It follows that we can never really explain everything 
exactly. Trying to stick to rigorous explanations from "base 
fundamental" principles results in not being able to explain anything 
at all. In practice there comes a level low down that has to be taken 
as a starting point without being explained. Take the "Big Bang" or 
God, for example. I think this is true for any discussion of 
anything.

CLARIFICATION:

By stressing the "VERY fundamental" I was implying that we can IN 
THEORY build a model of a brain. As Mark Weiser says, in practice this 
would also require us to duplicate the "software" of the brain. The 
operation of a neural network is defined by its structure and its 
"learning" capability by its update algorithms. The structure defines 
what a neural network can and cannot do to a much greater extent than 
the structure of a programmable electronic computer does. In a 
mathematical neural network, its "memory" is defined by the strength 
of interconnections between the various nodes, and the value of any 
feedback connections at a particular time (i.e. the momentary state of 
the system). Similarly, in a biological NN its memory is defined by 
the chemical strength of interconnection between the neurones, and the 
electrical feedback impulses at any one time.

The parallel with the electronic computer is that we would not only 
know how the circuits worked, but also what state each circuit was in; 
i.e. the memory contents of the computer, which are its software.

So, in theory, it is possible to look at an electronic computer and 
copy its circuits, and then copy the electric potential of each of 
the circuits (i.e. the software) and then we would have a copy! But 
in practice this would be virtually impossible to do exactly.

Similarly with the brain; if we copied the operation of each type of
neurone, the neural layout and interconnections, the momentary chemical
strength of each interconnection, and the momentary electrical state of
each neurone, then we would have a perfect replica -- and assuming (1) is
correct in my original article, then this is possible in theory. However,
if copying a computer exactly is virtually impossible in practice, then
this is surely much more so!! 

In my opinion, a more plausible scenario (in practice) for some time in
the next one hundred years, would be the following:

If we just simulated the operation of each type of neurone and the neural
layout and interconnections, then we would have something resembling a
BLANK human brain to some extent, depending on how good the copy was. We
could then start to teach it things, i.e. to build up its own chemical
interconnection strengths and electrical impulses.  The resulting brain
would be human in form but its behaviour would depend very much on which
sensory inputs we allowed it and what we taught it. To take the computer
analogy again, we wouldn't be giving it a replica of the "software" in
another brain; it would be creating its own by learning from its inputs. 


 Dean Pignon,  London, UK.

<Dean.Pignon@Cognos.COM>



--------------------------------------------------------------------
to (un)subscribe  the Forum just mail to
fleshfactor-request@aec.at (message text 'subscribe'/'unsubscribe')
send messages to fleshfactor@aec.at
--------------------------------------------------------------------


[FleshFactor] [subscribe]