Re: virus: Re: neural nets

David McFadzean (
Sun, 5 Jan 1997 17:08:38 -0700

> From: Alex Williams <>
> Date: Sunday, January 05, 1997 3:41 PM

> Replace Turing by Von Neumann in every instance I've used it over the
> course of this discussion for the past X days, except when I actually

Will do.

> meant Turing ... (Another discussion I'm in on another ML is
> centering about Turing tests, the only reason I can figure for
> brain-death of such magnitude.)

I would be interested in a summary of the Turing test discussion
if you are willing. (Which mailing list?)

> > Are you claiming that neural nets can't do computation, or that
> > computation takes place at a higher level, or something else?
> I'm suggesting that Von Neumann computational systems are inherently
> more computationally flexible than NNet architectures because the
> latter don't do `computation' as such by storing values in memory,
> doing ongoing comparisons to state changes, etc. NNets take in
> patterns and output patterns without `computation' going on between.

There are two ways we can evaluate this. As an implementor of artifical
neural networks I can assure you that computation is involved, after all
I wrote the code that does the computations (i.e. summing activation
levels over all inputs and mapping the result to an output function).

The other way is if the network is implemented in hardware or wetware,
in which case the results are produced by virtue of electrical circuits
rather than interpreting instructions. I would agree that there is no
computation taking place *at that level* but when interpreted at the
level of the network or higher, there may very well be computation.
To illustrate, imagine a flipflop device implemented as a 4 neuron
network (easily done), then using several of these flipflops to implement
a NAND gate (also very straightforward), the using the NAND gates to
build a computer.

> Even given the Mondo-Mega NNet that models an entire 4meg memory
> Pentium from state to state, it can't model the exact same machine
> with 5meg of memory. To do so requires an entirely new NNet with all
> attendent learning time and physical expansion.

Not if built in the way I just described.

> I agree with you, but we've had discussions over whether or not memes
> that don't exist within a given mind are memes (and you'll recall I'm
> a heretic there, at least). I have to question everything once in a
> while.

Allow me a question then: How is your "memesphere" different from
a belief system?

David McFadzean       
Memetic Engineer      
Church of Virus