Re: The So-Called "Failure" of 5th Generation and Logic
- From: "JAK" <jak@xxxxxxxxxxxxxxxx>
- Date: 12 Jan 2006 14:05:05 -0800
Perhaps the answer may be somewhere in-between. Significant
experimental results point to decision-making as a mimic of Herbert
Spencer's "Survival of the Fittest." Perceptions from the senses are
first used like "calc keys" for a memory "fetch". The best match of
memory to sensory input "wins" this first hurdle (though secondary
matches may be retained). This matching was described in 1959 by O.G.
Selfridge (Pandemonium Model) and in the 1990s by Dominic Masaro (Fuzzy
Logical Model of Perception - FLMP). It appears that behaviors are
linked to the stored memories. If the image perceived is round and
red, both a ball and an apple may be fetched from memory. Further
investigation (additional sensory input) will solidify the
identification. (It smells like an apple, so I guess it is one.)
Assuming an apple is the final determination, the associated behaviors
include eating or leaving alone. If the hypothalamus has raised the
level of hunger, this additional internal stimulus invigorates the "eat
apple" behavior, and eating the apple will likely ensue. However, if
the apple belongs to someone else, this aversive internal stimulus will
evoke a "leave it alone" behavior. The two compete within the nucleus
reticularis thalami (nRT) - the outer shell of the thalamic bulbs tied
bidirectionally to the hypothalamus to form the central components of
the diencephalon - the crucial component for mental activity. This
competition determines which behavior takes control of the organism
(see James Newman and Bernard J. Baars as well as Global Workspace
Theory):
"Newman, Baars and Cho reach a similar conclusion. They deduce that
the frontal lobes perform an 'executive function' over the limbic
system. The researchers take pains to point out the fact that this
executive function is not indispensable to consciousness - thalamic
competition is judged the more fundamental manifestation of conscious
life:
Note that an executive system is not an essential requirement for
consciousness. That this is not the case is illustrated by the
literature on extensive damage to the frontal lobes of the brain....
In terms of the [Global Workspace] model we have presented, it is not
executive attentional processes, but the selective binding of
coalitions of active cortical modules via a thalamocortical competition
which is the sine qua non for the generation of a coherent stream of
conscious representations."
- (http://mbdefault.org/8_identity/3.asp)
This strategic design lends itself readily to a control/feedback system
(see William T. Powers "Perceptual Control Theory) whereby exacting
programmatic logic is unnecessary. In fact, some success has been
garnered for robotic activities without a cpu or programmatic logic -
such as Ron Blue's Ricci
(http://library.thinkquest.org/18242/gather/cgi-bin/programs/3.html).
">There is no real clock per se, except for the purpose of comparing
input to
>memory at certain rates. It is analog. No instructions to follow. It
>learns by experience."
....
>The sensors are CDS cells. It allows LR (Little Ricci) to sense
>colors. It is known (Einstein won the nobel prize for this discovery)
>that different colors produce different energy levels in the cds cell.
>The "processor" generates different levels of voltage, *pure* voltage
>(IOW no current) in what we call a prism. This is an electronic prism.
>Each output varies a little different from the other. The "calculation"
>is done within what we call the neutronics chamber. The outputs are done
>within the prism splitter. The outputs are then input into what we call
>string memory. Each clock cycle stores the particular level along the
>string in the manner of a shift register. As the input remains the same,
>the memory is compared through a comparator to the splitter output and
>reinforced. Hence the learning. Hence the way our own brains work. All
>analog, all with no current, all quantum, since it is the quantum
>properties of electron charge waves and the crystalline structure of the
>neutronics chamber of which we are taking advantage. Parallel in
>Parellel due to the density of the crystal (silicon) atoms/electrons.
>
> Our brains really don't process information as a CPU, but retains
>memory in its 10 trillion neurons. Reinforced by practice and
>experience. We make decisions based on what we *remember*. We see a
>situation which looks familiar, we are simply comparing sensory input to
>memory. Correlational opposition. Then, we send signals to the
>appropriate body parts to take action.
>
> This is Little Ricci. But Little Ricci is more like an amoeba, or
>any other single cell animal. It exhibits the same behavior. The first
>prototype stayed away from bright sunlight. Single celled animals do
>also. It went towards a red wagon. Single celled animals go towards
>friendly environments to survive. We did not program LR to do this. It
>learned what it liked and didn't like all by itself! The difference
>between animals and us is that we have mucho more neurons and memory
>retention."
- (http://list.dprg.org/archive/1998-January/002970.html)
.
- Follow-Ups:
- Re: The So-Called "Failure" of 5th Generation and Logic
- From: mimo_545
- Re: The So-Called "Failure" of 5th Generation and Logic
- References:
- The So-Called "Failure" of 5th Generation and Logic
- From: markwh04
- Re: The So-Called "Failure" of 5th Generation and Logic
- From: Traveler
- The So-Called "Failure" of 5th Generation and Logic
- Prev by Date: Re: Why do memories die down over time?
- Next by Date: Re: Mechanical Dualism versus Naturalized Epistemology
- Previous by thread: Re: The So-Called "Failure" of 5th Generation and Logic
- Next by thread: Re: The So-Called "Failure" of 5th Generation and Logic
- Index(es):
Relevant Pages
|
