Date: Thu Oct 20 2005 - 00:22:34 MDT
His ideas are well-done and useful (which doesn't mean that they are
correct, of course).
His book On Intelligence is well worth reading. Here's my 30 second
- The cool part of human intelligence happens in the neocortex.
- The neocortex is all uniform: every module is functionally equivalent to
- The purpose of the neocortex is to do pattern matching.
- Intelligence happens when you stack many pattern matchers together in a
Further details that he considers important are that information travels
down the hierarchy as much as up, that time is an important player in the
system, and that prediction is a very important role of the pattern
Jeff recently started a company (Numenta, not to be confused with Novamente
:-) to build the core technology that he believes is involved. They demoed
at AC2005, but didn't really have enough built to allow one to make an
informed judgment about how they're doing.
My personal read is that Jeff's a super-smart guy with talent, ambition, and
money, as well as some good ideas. I get the sense (possibly unfairly) that
he's a little simplistic in his approach: to my mind, he comes across as
somewhat dogmatic that almost everything interesting is the result of this
single absolutely invariant module.
I've recently started work on a visual system which is in many ways similar
to Numenta's, although I'm inclined to stray further from biological
faithfulness than they seem to be. I'll probably have a more informed
opinion about how useful the approach is in a few months...
From: firstname.lastname@example.org [mailto:email@example.com] On Behalf Of Chris Capel
Sent: Wednesday, October 19, 2005 4:56 PM
Subject: Jeff Hawkins' theory of cognition?
Does anyone know if his ideas are well-done and useful?
-- "What is it like to be a bat? What is it like to bat a bee? What is it like to be a bee being batted? What is it like to be a batted bee?" -- The Mind's I (Hofstadter, Dennet)
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:52 MDT