From: Brian Atkins (
Date: Wed Jul 04 2001 - 00:21:50 MDT

Christopher McKinstry wrote:
> A way of thinking about the whole Mindpixel approach is to think of
> signal processing: Think of complex thought as being like a complex wave
> form and binary propositions being like component sine waves - you even
> get to use the same math, which is very, very cool. And quite telling as
> well.

I don't think anyone is disputing that you can train a neural network
to respond "correctly" to all of your binary propositions. Just like
you can train a neural network to recognize specific words in speech.
But in neither case do I see how that leads to the neural network
actually understanding anything, or being able to pass the Turing test.
Even if you get your trained neural network to be able to interpolate
or otherwise get right questions that it hasn't exact data for, I don't
see that leading to any real AI. At most what you get is some kind of
advanced search engine. Are you just hoping the neural network will
magically turn into something like a human mind? What am I missing?

Brian Atkins
Director, Singularity Institute for Artificial Intelligence

This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:39:41 MDT