Robot Dreams

George Bajszar gyuri at
Mon Sep 11 02:49:00 EST 2000

In article <8pi14q$grl$1 at>,
  George Bajszar <gyuri at> wrote:
> > > There are reasons people sleep. We know if we don't get enough
> > > sleep, our short and long term memory is stronly affected, so
> > > it is a natural conclusion to think that a cleanup or
> > > reorganization of our memory is occurring during sleep, more
> > > particularly the dreaming process is strongly associated to
> > > memory. For an AI system such cleanup could be setup daily,
> > > or simultaneously while it is awake.
> > I call this mental "garbage collection", without which our
> > brains would become full and the stack will get overrun.
> > Insanity then ensues.
> >
> > IMO,
> > DLC
> Everything that we "pay attention to" will get recorded into our
> memory. This has been well demonstrated by various neuropsychological
> experiments.
> In one test called the Recall Test, 15 cards with various images or
> words are flashed, and than the subject is asked to recall the images
> or words presented, regardless of their original order of
> Only 6 or so on average is recalled. That does not mean that the
> of all 15 cards are not stored in the brain.
> In another test called the Recognition Test, the 15 cards are flashed
> to the subject and then those cards are shuffled in with another set
> unseen 15 cards. The 30 cards then are flashed one by one and the
> subject is asked to identify which cards he has already seen. On
> average, 97% of the cards were correctly recognized, even when pausing
> for over 20 minute delay after presenting the first 15 cards.
> That shows that the brain is like a camcorder. Everything is stored
> that we pay attention to. If everything would be stored, our brain
> would certainly become full quite soon. So special filtering is needed
> to filter out useful information to keep, and clean out the unuseful
> ones so next Morning we wake up with a fresh clean mind. Dreaming
> have such a major purpose.
> Supermemory techniques also demonstrate that everything gets recorded.
> Techniques to access data are used by supermemory experts by building
> up associative links to them.
> Perhaps Schizophrenia is a result of an overloaded brain. When the
> brain gets full, new memory gets written out of control onto areas
> where useful memory is already stored. The stable neural network
> becomes unstable as the wiring of neural networks will begin to deform
> into unstable randomness, and neurons begin sending signals randomly,
> resulting in hallucinative effects. That is my guess.
> George Bajszar

Perhaps sleeping has yet another purpose. I just came up with this
small theory:

As the day goes by, our neural nets storing memory are forming with
data. But as I already described, a lot of memory that is unneeded is
also stored during the day into random neurons. At the end of the day
we go to sleep, and the brain if full with garbage data. As we go to
sleep, the brain slowly shuts off into deep sleep mode. All this is
controlled by chemical processes.

Then from the deep sleep mode slowly arises a less deep sleep state
where certain neurons wake up. Neurons that are in a stable network
representing strong connections remain asleep. The chemical nature of
strong connections allows this to take place.

However unstable neurons that did not form significant network
connections during the day will begin in this state to fire random
signals. This random process sparks the dreaming process.

The purpose of this would be for the brain to identify garbage neurons
and reset them. Maybe that is why we don't remember dreams typically
because those neurons that sparked those responses that lead to the
dream were cleaned up for good from any further signaling.

George Bajszar

Sent via
Before you buy.

More information about the Neur-sci mailing list