Sergio,

On Fri, Jun 29, 2012 at 12:34 PM, Sergio Pissanetzky <[email protected]
> wrote:

> > there is SO much more going on there … it will be possible to infer much
> of what we cannot see, and to substitute things that will self-organize to
> fill in for what we can't infer. ****
>
> SERGIO > I prefer to view that "SO much" as consisting of two parts. One
> that we can diagram, image, simulate, understand. This part can, to a
> point, be modelled on a computer, and yes, it is very complex. The other
> part we can not understand. It is the "infer" part, the "self-organize"
> part.
>

We will be able to see a LOT more in living neurons once scanning UV
fluorescence microscopes become available. Now, there are now in common use
non-scanning UV fluorescence microscopes that show the chemistry, but they
only provide low-resolution 2-D images because they cannot separate out the
3rd dimension. We simply don't know what we can see with better equipment
that provides true 3-D images with full UV resolution. I suspect that we
won't be able to see everything, but that we will be able to see enough to
see how things work at a functional level.

In short, EVERYTHING is now hung up because no one wants to invest a few
million dollars into designing some new equipment.

I think you and I are saying essentially the same thing. Your thinking is
> different, you propose to "infer," meaning we, humans, doing the inference.
> I want the machine to do it.
>

We humans will probably have to figure out that when you see certain things
in the images, it means that predictable hidden things are also happening.
We then program this knowledge into the diagrammer, whereupon the
diagrammer forever after takes care of those things.

> ****
>
> ** **
>
> > hypothalamus****
>
> Sergio > I did read the website time ago, but I keep this material in a
> queue. My approach is constructivist, bottom-up, and I am not yet ready to
> move that much higher. For brain-storming, the hypothalamus is ideal. But I
> am ready for a model of the retina,
>

The retina is IDEAL because it learns its functionality simply by observing
its input. Most people concentrate on the particular things being
recognized by the retina. The "baby" is in how it self-organizes to achieve
its functionality, and the "bathwater" is its resulting functionality.
Nearly all research throws the baby away and studies the bathwater.

not a fully-functional one but a minimalistic one that is simple enough for
> understanding and processing yet advanced enough to teach us something
> useful and help us in the right direction for the rest. If this model
> succeeds in simulating some parameter that has been observed in an actual
> retina (compression?),
>

Demonstrating unsupervised self-organization would be SPECTACULAR. Without
that, everything else is unsupportable speculation.

but can not be explained by other means,
>

There are ALWAYS a dozen or so bogus explanations for just about
everything. THAT is why most experiments fail.


> then one would have a nice tale to tell the wet-lab people.
>

NOT until you first get their attention. THAT can be quite a challenge.

Steve



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968
Powered by Listbox: http://www.listbox.com

Reply via email to