[keghn]

HTM School Episode 0: HTM Overview:




HTM School Episode 1: Bit Arrays:

! No longer available




HTM School Episode 2: SDR Capacity & Comparison:

[Korrelan]

Cool videos again.

I had always thought Numenta & I were working along similar lines, but apparently not lol.  I never went too deep into understanding their technique as I didn't want to cloud my own thinking (I'm getting old lol). I have now read up on their architecture, interesting.

I’ll look forward to the other videos.

[keghn]

https://groups.google.com/forum/#!topic/artificial-general-intelligence/UVUZ93Zep6Y

[djchapm]

I really dig the diagram of the neocortex - very cool.  After seeing the video, I start to feel they are jumping the gun in trying to use this HTM network/structure to produce results with things like GPS and Traveling.  And I wasn't in agreement with the parallel he drew to deep learning where the first layer identifies edges, second identifies shapes, etc.  I do agree with the insight that last outputs have to be are part of the input. 

What about his assigning feeling, emotion, love, hunting, etc all to the 'old' parts of the brain underneath the neocortex, and saying that those things are all common in all mammals and not related to intelligence.  He says reptiles haven't developed this neocortex and thus not intelligent, but then he goes to describe how the neocortex is responsible for seeing and classifying what is seen similar to a deep NN.  Reptiles do this - so must not be part of the neocortex? 

Not trying to pick apart... I just think that based on how they described HTM, their own examples are contradictory to what they say it is used for.

DJ

[keghn]

 Numenta has interesting stuff. I just like to have a compare to my AGI theories:

https://groups.google.com/forum/#!forum/artificial-general-intelligence

[keghn]

CLAASIC: a Cortex-Inspired Hardware Accelerator:

https://arxiv.org/abs/1604.05897

[8pla.net]

Recently, I read something recommended by Art on this site.  So, that inspired me
to look into it a little more.  Crows, (which are birds) have the reasoning skills of a
seven year old human child.  Crows use different tools and so forth, which is quite
amazing to watch on video.  So, if birds do not have a neocortex, then what does
that say about reasoning skills?   Is it possible that reasoning skills are old brain?

[keghn]

One Way That Bird Brains Could Be Superior to Mammal Brains:
http://io9.gizmodo.com/5948169/one-way-that-bird-brains-could-be-superior-to-mammal-brains


The bigger the animal, the better to push away competition for a mate.
The bigger the animal the slower it will evolve.
To be able to fly, a animal need to be smaller than non flying animals. Giving the birds a evolving
advantage.
 A lighter more nimble male bird is better at pushing other male away, than bigger. Like
the crow and mocking bird or sparrow.

 Bats are under the same conditions too, but use it to become fast evolvers against viruses,
except for the bat that hibernate. Hibernation is evolution on hold.

 Insects too, and non flying insects that do not have lungs, limit their size.

 And then there comparing of the crow brain against other birds.

[madmax]

I think that the hippocampal region of the mammals is similar to the birds brain but is somehow evolved differently.Birds are moving very fast so they need to perceive surrounding much better.So they need like human cortex memory but without hierarchy together with hippocampal spatial memory so they together make some sort of hierarchy.They have like photographically memory,this you can observe in their ability to mimic various sounds.But they dont reasoning human like.

And the neurons are similar to the cortex ones so HTM in my opinion could apply here to.
And in my opinion human ability for cognition is not only about cortex.