A Theory of Sleep

Ed van der Meulen ameulen at users.sourceforge.net
Tue Aug 14 17:41:51 EST 2001


"Arthur T. Murray" <uj797 at victoria.tc.ca> schreef in bericht
news:3b77fb27 at news.victoria.tc.ca...
> "yan king yin" wrote on Mon, 13 Aug 2001:
>
> > Hi All,
> >
> > This is the first draft of my paper.
> > Critiques or comments welcome.
> >
> > The HTML version with a few figures is at:
> > http://www.angelfire.com/myband/sevenless/Sleep.htm
> >
> > =================
> > A Theory of Sleep
> > =================
> >
> > The Generative theory of sleep is proposed here.
> > It is based on synaptic selectionism, summarized below.
> > The new theory construes REM sleep as the generator of
> > "random" synaptic connections; and slow-wave sleep as
> > the stage where inadaptive synapses are eliminated.

> http://sourceforge.net/projects/mind/ in the Docs area.

> > (13 August 2001) contact author: y.k.y at lycos.com

Sleep functions are in my view important for all kinds of
neural nets, mainly dealing with it's inner unrest.

On the site http://nnw.sourceforge.net you can find
demos of neural nets having a few sleep functions. And if
you know how important categorizing is for our human
intelligence, you can't underestimate these sleep function.

But randomize synaptic connections is for me too wild.
All happens with all kinds of constraints. In my models
I certainly do use variation with some amount of chaos,
which is inherent in the signals and some in the neurons.

Internal generated signals have indeed some chaos, but
the overall behavior is fairly logical.

Important also is the large difference in the size scales,
between a single connection and overall behavior,
which I see as emerging with often unpredictable
behavior, however frequently within bounds.

I tell you the clue of my view. It's very simple in my project,
look at a level of detail (LOD) or scale where you don't see
the individual neurons anymore, but only streams. Those
streams can stream forward and sometimes backward.,
with only a few extra features. The number of degrees of
freedom is low on this LOD. We have coded those
streams, that simple. But we've lost the knowledge that our
cell is the representant of one real neuron, it could stand
for a cluster of cells, but with the amazing bottom-up
functioning.

Bye,
Ed





More information about the Neur-sci mailing list