Re: Biological Neural Nets, ANNs, and Turing Machines

From: Roger L. Bagula (rlbtftn_at_netscape.net)
Date: 07/21/04 

Date: Wed, 21 Jul 2004 22:20:44 GMT

You should try reading
" A New Kind of Science"
by Steven Wolfram
He goes into such problems in detail
as related to cellular automata
( which in turn can behave very much like neural networks).
I think that you can emulate a Turing machine with an Neural network,
but it is very much the dumbing down of a much more complex learning
process.
The point of biological neural networks is that they
find solutions missed by straight Turing logic machines
due to a stochastic effect/chaotic effect.
Genetic algorithms are another modern area where
the attempt is to get at self-organizational effects of biological
systems.
I think maybe Per Bak's later work on sand piles might help your
thought processes to become simpler.
Sina Tootoonian wrote:
> Hello all,
>
> I was wondering if those of you wiser than myself may be able to clear
> up an issue that has been nagging at the back of my mind for some
> time:
>
> Is there any fundamental difference between the types of computation
> that biological neural networks perform and those that a universal
> turing machine can? I.e., if a biological neural net is studied well
> enough, it can be modelled as an ANN, which are run on ordinary PCs,
> therefore, in principle, the entire brain can be programmed in C. Is
> this logic correct or am I missing something here? And I don't mean
> implementation details. I understand that to actually implement any
> such system, the hardware on which the software runs may have to use
> something other than von-Neumann architecture (e.g. Connection
> Machine). I am thinking of issues analogous to what happens in
> deterministic dynamic systems, which can theoretically be predicted as
> accurately as required, but in practice exhibit chaotic behavior.
>
> Any insights would be greatly appreciated.
>
> Cheers!
>
> Sina 

-- 
Respectfully, Roger L. Bagula
tftn@earthlink.net, 11759Waterhill Road, Lakeside,Ca 92040-2905,tel: 
619-5610814 :
URL :  http://home.earthlink.net/~tftn
URL :  http://victorian.fortunecity.com/carmelita/435/

Relevant Pages

  • Re: Biological Neural Nets, ANNs, and Turing Machines
    ... as related to cellular automata ... (which in turn can behave very much like neural networks). ... I think that you can emulate a Turing machine with an Neural network, ... Sina Tootoonian wrote: ...
    (sci.nonlinear)
  • Re: Can we neural networks to predict code change?
    ... If you mean a form of neural networks that can be implemented on any ordinary computer, then you are wrong; neural networks can't do anything that's not possible with a Turing machine. ... it's clear that neither the human brain nor any kind of computer can give an exact answer in the general case. ... Your opinion that the ANN could likely approximate some of the functions of the human brain seems uncontroversial. ...
    (comp.theory)
  • Re: Can we neural networks to predict code change?
    ... If you mean a form of neural networks that can be implemented on any ... that's not possible with a Turing machine. ... large code base - with reasonable accuracy. ... is essentially what the human brain does, ...
    (comp.theory)
  • Re: Can we neural networks to predict code change?
    ... If you mean a form of neural networks that can be implemented on any ... that's not possible with a Turing machine. ... large code base - with reasonable accuracy. ... is essentially what the human brain does, ...
    (comp.theory)