Great question.  You are the first person to ask this.  I don't have time to
go through each of the predictions in the book in detail, but I will share
some general observations.

 

- Lots of neuroscientists thought the book was great and novel.  Some were
critical as they felt what I wrote about was not new.  No one objected to
what I wrote based on content, at least I am not aware of any.

 

- When the book came out I spoke to a few neuroscientists who had read the
book and had specific knowledge about some of the predictions.  These
conversations were a bit inconclusive, such as "there is some evidence in
support but it isn't clear".  No one said "yes you got this right", or  "no
you got this wrong", or "here is contradictory evidence".  That was a little
disappointing.

 

- Many of the predictions in the book involved details of physiology and
anatomy at a fine level of specificity.  There has been relatively little
progress made on tools that can test these predictions.  Many of the funds
that in the past went to anatomy and physiology have been directed to fMRI
and other "hot" areas.  So these kinds of in vivo cellular studies have
become a bit of a back water.  For example the number of people in the world
who study connections in layer 1 could probably be counted on one hand and
it may be as little as zero.  Bottom line is there are few tools and few
people who are capable of testing these predictions.

 

- In neuroscience there isn't a history of theorists and experimentalists as
you see in physics.  Neuroscience experiments are hard, expensive, and can
take years to complete.  Because of this most experimentalists don't test
other people's theories and often don't even share their data.  So I wasn't
expecting people to rush out and design experiments to test these
predictions, and they didn't.

 

Finally, I have not been out scouring the literature for verification of
these predictions.  There may be supporting or contrary evidence that I am
not aware of.

Jeff 

 

 

 

From: nupic [mailto:[email protected]] On Behalf Of Oreste
Villa
Sent: Wednesday, September 04, 2013 8:41 AM
To: NuPIC general mailing list.
Subject: [nupic-dev] "On Intelligence" Appendix predictions..

 

I remember reading "On Intelligence" the first time in 2006 (then in 2009
and then few months ago). 

I am wondering which predictions have been confirmed or denied since the
book was first published in 2004.

It is almost 10 years (I know time flies when you are having fun :-) ) and I
am sure there must be something new out there with respect to these
predictions.

 

I know literature is huge, but is there someone up to date that is willing
to share some knowledge?  

 

Example of predictions, for people that did not read the book (not to
advertise but you should read it at least once): 

 

Prediction 4

One class of cells in layers 2 and 3 should preferentially receive input
from layer 6 cells in higher cortical regions.

 

Prediction 8

Sudden understanding should result in a precise cascading of predictive
activity that flows down the cortical hierarchy.

 

(there are 11 predictions in the appendix and other through the all book)

 

Thanks,

_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org

Reply via email to