On 13/06/06, Yan King Yin <[EMAIL PROTECTED]> wrote:
One question is whether there is some definite advantage to using NNs
instead of say, predicate logic. Can you give an example of a thought, or a
line of inference, etc, that the NN-type representation is particularly
suited?
The distributed nature of computation in a Neural Net allows the
potential for logical connections such as SectionA of neural tissue is
using too much energy non-usefully, therefore its energy supply should
be decreased via neuro vascular coupling. That is as computation is
embodied and localised there is the potential for monitoring and
control.
Predicate logic is great for representing the external world, it can
even represent the internal state of the machine but without the
correct hooks it cannot control the inner state. Could peoples designs
for logical AGI meaningfully follow the command, "Clear your mind". I
know backprop can't do this either, but backprop is still very much
focused about "learning about the world", rather than "learning what
should be done".
Will Pearson
-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
