Robot Dreams

George Bajszar gyuri at usa.net
Thu Sep 14 23:50:37 EST 2000


In article <hq5w5.82708$UO.258456 at news22.bellglobal.com>,
  "Rick Craik" <NOSPAMrcraik at ntl.sympatico.ca> wrote:
>
> George Bajszar wrote in message <8pq1o7$uq$1 at nnrp1.deja.com>...
> >In article <39BFF09B.1FAC at vic.ozland.net.au>,
> >  jgcasey at vic.ozland.net.au wrote:
> [snip]
> >Perhaps what is happening is that neurons generate
> >random noise with their signals in the background of
> >our attention,
>
> Just a comment:
>
> I can't resist the temptation to paraphrase Albert
> Einstein with: "Neurons do not roll the dice."
> But perhaps you are correct, the concept
> seems to hold a place in your reasoning.
>
> What other ways of generating signals that appear
> random do we have? How about "chaotic", "arbitrary",
> or "out of tune" for examples.
>
> "Arbitrary" could be used instead of random to
> indicate that the neuron decided, or arbitrated,
> when and what signals are fired off. I would suspect
> that randomness is filtered or tuned out at the neuron
> level. The nature of neurons seems to be to make
> decisions.
>
> "Chaotic" decisions might be reasonable to assume.
> Arbitrary inputs are processed depending upon
> the current state of the neuron, a state which may
> be unpredictable due to some external influence.
>
> The "out of tune" aspects, personally I would like to
> explore . Neurons appear to have something like
> a digital to analog to digital signal processing
> capability. This type of exploration could make
> some sense out of a chaotic signaling neuron.
>
> >and often the random combination of
> >messages spark the seeked correct answer.
>
> If I had to design a neuron system, I'd try to
> leave the random combinations outside
> in the environment. Those outer neurons
> have spent a lot of years evolving to get
> rid of them. :)
>
> Regards,
> Rick

Each neuron is an independent little machinery. They must have the
ability to function independently to fit into the network of an entire
system composed of other identical neurons. They can make their own
decisions on what other neurons they must connect to. When neurons
connect, they establish a working partnership. All these abilities are
carried by each neuron independently.

In general if an independent component that can fit or blend naturally
into a system of other already functioning components, it must be
harmonious, synchronized with the other components. The smooth
harmonious functionality of a working system is the "music" of the
system.

Perhaps in a way that is what you were referring to in the information
processing level.

I also think that information received through sensory inputs might be
represented as unique waves:

Perhaps the memories storing the image of an apple might represent
similar brainwaves to the brainwaves sparked by our eyes when seeing an
actual apple. Perhaps we can visualize an apple in our mind because the
brainwaves of the memory of the apple's image sparks the same waves and
those waves are received by the same neurons whose job is to transmit
images from our eyes into our brain.

I think the theory I described above is quite old, it makes sense but
perhaps noone really knows if that corresponds with the reality.
Experiments have been made to record brain waves received during visual
inputs and were retransmit back somehow to the brain to see if those
images would reappear in our consciousness. Those particular experiment
failed.

We may observe waves in the brain, but as far as information is
concerned, the reality may be more digital in nature. What may be
important or relevant is individual neuron level activity. Brain waves
may be just a sideeffect we are observing of repetative neuron
activities. The image from the eye's retina cells are transmitted
individually to neurons where the information moves from neuron to
neuron, and is processed maybe independently. But then again, it may be
otherwise, I don't know.

George Bajszar


Sent via Deja.com http://www.deja.com/
Before you buy.






More information about the Neur-sci mailing list