WHITE PAPER: Numenta's Hierarchical Temporal Memory (Jeff Hawkins and Dileep George)

From: Neil H. (neuronexmachina@gmail.com)
Date: Thu May 18 2006 - 13:30:36 MDT

FYI, it looks like Jeff Hawkins and Dileep George have made available
a white paper on Hierarchical Temporal Memory, the neurally-inspired
framework discussed in Hawkins' book, "On Intelligence."

I'm still not convinced about how well their system will work on more
natural/complex stimuli, but it's an interesting read nonetheless.

---------- Forwarded message ----------
From: Numenta <noreply@numenta.com>
Date: May 17, 2006 1:49 PM
Subject: Numenta Newsletter #3
To: neuronexmachina@gmail.com

(Please note: you may be receiving this email as a duplicate of an
earlier email. We're sending the newsletter to some subscribers a
second time because some of the links did not work in the earlier
email. We apologize for the inconvenience!)

May 17, 2006

Dear Newsletter Subscriber:

This is the third newsletter from Numenta.

We are pleased to announce that we have completed a first white paper
on Hierarchical Temporal Memory. It is now available on our web site
http://www.numenta.com/Numenta_HTM_Concepts.pdf. The white paper
introduces the concepts and theory behind HTM technology. For those of
you who would like to learn more about Numenta's technology, this is a
good place to start.

In order to create a stronger product offering, and to facilitate more
targeted communications, we ask you to tell us a little bit about your
interest in Numenta. Please follow the profile link at the bottom of
the email in order to answer two quick questions. (Doing so is
entirely optional, and failure to do so will NOT remove you from our
list, but we'd appreciate it if you would take a minute to respond to
our two questions!)

We continue to make good progress on designing and developing our
Hierarchical Temporal Memory (HTM) toolset. We have completed a first
set of learning algorithms, and are moving towards testing them in our
production environment. We recently decided to move from the MatLab
development environment for the learning algorithms to using Python,
which will enable more rapid prototyping and a smoother integration
into the C++ production environment.

The overall architecture of the system has continued to progress,
along with early work on the tools themselves. Our hope is to complete
an integration milestone, bringing together the learning algorithms
within the system architecture, within the next month.

We have three partners participating with us in our Numenta Partner
Program today. I'm pleased to say that we have hired Phillip Shoemaker
to be our Director of Developer Programs. Phillip has substantial
experience in working with developer communities.

Our tentative plan remains to release a beta version of our toolset
before the end of 2006. It is likely that our first developer
symposium will be in 2007, giving developers a chance to work with the
tools for some time before meeting together.

Thank you for your continuing interest and your support. We are very
pleased with our progress, and continue to feel that HTM technology
offers the potential to solve some significant problems in computing
across a broad range of application areas.

Donna Dubinsky
CEO, Numenta

Our first two newsletters can be accessed at our web site,


To unsubscribe from this newsletter, follow this link.

You may update your profile by following this link.

Copyright 2006, Numenta, Inc. All rights reserved.
All materials contained in this newsletter are protected by United
States and international copyright laws and may not be reproduced,
distributed, transmitted, displayed, published or broadcast without
the prior written permission of Numenta, Inc. You may not alter or
remove any trademark, copyright or other notice from copies of this

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT