• Members of the previous forum can retrieve their temporary password here, (login and check your PM).

Jeff Hawkins on Intelligence: A hierachical model of the brain

Migrated topic.
embracethevoid

embracethevoid

Rising Star

Five part video:

Absolute must watch for anyone interested in how intelligence operates. He theorises/observes that the brain uses hierarchichal processing of sensory input to produce perception & intelligence. It is said that the all the parts of the hierarchy actually use the same processing program to carry out their respective functions: the specialisation arises from the fact that different pathways are connected to different organs. Seems like using 3000 instances of Microsoft Paint to draw the Mona Lisa and actually doing it well!
 
..yeah cool embracethevoid, i posted the same link the other week..:)

the 'tree' structure heirarchy is really groovy, and that all neurons perform the same function..pattern recognition..
..he doesn't have a lot to say about the 'upper level' of the heirarchy though..?
.
 
nen888 said:
..yeah cool embracethevoid, i posted the same link the other week..:)

the 'tree' structure heirarchy is really groovy, and that all neurons perform the same function..pattern recognition..
..he doesn't have a lot to say about the 'upper level' of the heirarchy though..?
.
Click to expand...

Sure he does. He said that all the sections are pretty much identical. So his description of the lower levels in the hierarchy applies to the upper levels as well.

The lowest level to the sensory input is looking for immediate patterns, and the higher up you go, the larger the pattern. Each layer up detecting patterns of the patterns in the level below it.

Layer 1 - recognizes A, then B, then C, then D, then E, then F
Layer 2 - Recognizes ABC as a pattern, and DEF as a patern.
Layer 3 - Recognizes ABCDEF as 1 pattern, and predicts GHI will be next and passes that back to layer 2
Layer 2 - With information to expect GHI, it passes to layer 1 an expectation of G
Layer 1 - Expects G to be next.

That's the process as I interpreted it, with many more layers obviously, and with more than one Layer 1 inputs. Patterns are detected, then predictions are made.
 
..he's only looking at the cortex..there is input and feedback from the 'emotional' parts of the brain as well..the key questions to me are:
what makes the decisions?
what level of 'hyper-sequence' is the top level? i.e. does it have a 'universal' conceptualization..?

what is the self? (is where this is leading)
.
 
nen888 said:
..he's only looking at the cortex..there is input and feedback from the 'emotional' parts of the brain as well..the key questions to me are:
what makes the decisions?
Click to expand...

This is essentially the question I want to find the answer to.

It seems we understand how learning/pattern recognition/prediction works in the Neo-cortex (not fully, but to a degree).

There are still plenty of functionality in the rest of the brain, that makes use of this data, or acts on it.
 
I think that the same model would also aply to emotional 'data'. Within the neo-cortex, decissions are made that also involve emotions, so the information processing that involves emotions has to be compatible with the other processes.
 
Back
Top Bottom