Speculation is always fantasy, which is why we don't believe in what
we speculate. We believe in experimental facts, but not the theory
that told us to run the experiment.
> > > Just because Einstein did it, doesn't mean it is the right thing to
> > > do. Einstein also committed adultery, is that alright now too? Too
> > > bad Einstein didn't believe in God or that would be OK also.
>
> > It was 'right' because it was useful. It resulted in the theory of
> > general relativity.
>
> Christianity was useful. Communnism is still useful. Dictatorship
> is still useful. I see alot of uncertainty in saying that something is
> "right" because it is useful.
Einstein used pure speculation to advance science dramatically. It was
'useful in a scientific sort of way'. Moreover, it was not harmful in
any way whatsoever. The other things you mentioned cannot say the same.
> > I'm confused. The article you referred to talked about humans
> > developing a new brain mechanism.
>
> Not humans. Animals that were precurssors to humans.
OK.
> > If a neural network "develops"
> > (for lack of a better term) a mechanism that appears identical,
> > then what word should I use other than "develop"?
>
> No computer program has a *need* to assimulate or process
> memories "offline", which is what a dream is. That is what I
> understood the research on that dream to mean, and that is
> why I strongly reccommended that no one take my word for
> it but read that article too. Another perspective on this matter
> would help me alleviate my doubts about my understanding.
I agree that 'need' could be going too far. I guess I'm mainly
interested simply because this thing is duplicating the mechanics of
dreaming. (And because I don't like coincidences) It may be mere
coincidence, but I think it would be easily testable. See below.
> > If we don't
> > use "develop" for either the neural net or the human, the fact
> > still remains that the neural net is doing something that appears
> > identical to what the human brain does when it dreams. I claim
> > that this is an interesting result. I do not claim that I under-
> > stand how the neural net "developed" such a behavior, but I do
> > claim that it is interesting that it behaved in such a way.
> > Are you at all interested in the result?
>
> No, I am not. The fact that the "research" implied that dreams served
> no purpose and were random, goes contrary to psychological
> research on the subject, starting first with Freud's discovery of the
> hidden subconscious revelations behind dreaming.
Yes, but when Freud analysed humans, there was more than simply the
physical brain: there was sensory input hooked up, and there was an
intelligent component - and Freud based his conclusions on what this
intelligent component told him. This is why I would be interested
in seeing the following experiment:
Take that same neural network, and give it the ability to 'sense' some
environment, and also give the ability to communicate to some simple
queries. Then, allow it to go into its 'dreamlike' state for a while,
and finally, ask it some questions to find out how it interprets its
own flurry of activity. I would be interested in seeing what it says.
- JPSchneider
- schneids@centuryinter.net
(mailing from home this time)