Re: What did that thread indicate?


Traveler wrote:


> You cannot have a single homogeneous net that does so many things at
> once. Besides, some of these things must be done in stages. IOW, you
> must start with signal separation (one principle) and then do signal
> fusion (another principle). Only then can you do sequence formation
> and motor control. You cannot do these things in the same network. You
> must have multiple subnetworks feeding signals to one another. On top
> of that, you must do concept formation, conflict detection, motor
> coordination, reinforement learning, etc... There is a reason that the
> brain has multiple integrated cell assemblies each with its own
> function and principle. In the brain, bi-directional signal pathways
> between subnetworks are the norm, not the exception.
>
> You come out sounding like the nodes in your net can do all these
> things simultaneously. This is complete crap, Curt. You're dreaming
> and you're only fooling yourself.


We've been round this block a number of times with Curt over the past
year and more. He continues to hold the behaviorist dream that the
brain is some kind of tabula rasa, and all of the requisite systems for
intelligence will magically self-organize from blank white nada, based
upon repetitive input from the outside world. But I think, in fact, he
is slowly learning that he needs to build a lot of "prior knowledge"
[IOW, genetic specification] into the network in order to get it to do
anything non-trivial. I think the best example of this is the addition
of feedback loops within the net. Curt never answered when I asked him
whether these loops were able to self-organize in his net, without him
having to specifically build them into the architecture. Also, when the
topic of feedback began to be discussed routinely around here, I
remember Longley the base behaviorist having many shitfits, as was his
nature regarding anything that didn't agree with cult rules.

However, I think that, when the story is finally written, it will be
found that feedback loops are "the" key aspect to getting real brains
to have real "brains". And from a philosophical perspective, this is
also interesting. Way back when ... simple organisms with very simple
nervous systems [ie, no internal FB-loops] routinely interacted with
their environments via "external" FB-loops. Even single-cell amoeba do
this. They poke around, and respond adversely when encountering bad
chemicals in the water, etc. Later on in evolution, the organisms
started incorporating FB-loops "internally", first by connecting
sensory to motor [reflex arcs], and later loops within loops within
loops, as the brains continued to evolve. Now, we have the massive
white matter in our brains, which mostly carries FB-information for
comms between various brain areas. Nature discovered how to build
intelligence. And I think Curt is following this path too, once he sees
it.

.

Quantcast