From: Charles D Hixson (charleshixsn@earthlink.net)
Date: Wed Jan 18 2006 - 16:45:02 MST
On Tuesday 17 January 2006 08:24 pm, Mikko Särelä wrote:
>...
> Neural networks are equivalent with Turing machines. Modern computers are
> equivalent with Turing machines. Our brains run software that nobody thus
> far has replicated with modern computers. Do you claim that if one did so,
> the outcome would not act like a human mind and would not have
> consciousness? (If it looks like a duck and quacks like a duck, what is
> it?)
>...
Be very careful here. Neural networks is an ill-defined term. The
abstractions used in computer science may be equivalent to Turing Machines,
and there may be no more powerful computer representable with math (which all
seems true). But the actual neural networks aren't closely modeled by the
computer science neural networks. We have reason to believe that the actual
neural networks aren't more powerful than a Turing Machine, but we don't
actually know this.
You hypothesis (and mine) I find convincing, but it isn't proof. That awaits
the experiment.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:55 MDT