Papers and videos

This section contains papers and talks describing details of HTM theory.

HTM Cortical Learning Algorithms

Hierarchical Temporal Memory including HTM Cortical Learning Algorithms (PDF)
This document describes in detail HTM technology and the new algorithms for learning and prediction. This document will be updated periodically with additional material. Questions and comments regarding the new algorithms may be posted on the HTM Theory forum. Last Updated: September 12, 2011.

The following translations were done under a translation license from Numenta. Numenta has not verified or endorsed these translations, and advises readers not to rely solely on the translation, but to refer to the English original version, particularly for difficult concepts. For information on licensing rights for a translation, please see here.

Brains, Minds and Machines (video)
MIT recently held a symposium on Brain, Minds and Machines as part of their 150th anniversary celebration. Jeff Hawkins was invited to participate on a panel that addressed the question: Is it time to try again to understand the brain and engineer the mind? Jeff presents his answer to this question in this 10-minute video. This talk was given on May 4, 2011.

Advances in modeling neocortex and its impact on machine intelligence (video)
Jeff Hawkins presents the new HTM algorithms at the Beckman Institute, University of Illinois at Urbana-Champaign. This talk was given on November 12, 2010.

Advances in modeling neocortex and its impact on machine intelligence (video)
Jeff Hawkins presents the new HTM algorithms for a graduate class on neural computation at the University of California at Berkeley. This talk is very similar to the Beckman Institute talk above, but might be useful for those viewers who want to hear the talk given again with some shades of difference. In addition, this talk includes a question & answer section. Note that the video is not high quality. This talk was given on December 2, 2010.

On Intelligence
Jeff Hawkins first described his theory of Hierarchical Temporal Memory in this book. It has been translated into several languages.

Archives

The following materials have been superseded by new documentation.

Whitepapers

HTM: Concepts, Theory and Terminology (PDF)
This whitepaper goes into an in-depth discussion about the theory of Hierarchical Temporal Memory (HTM). HTM is a new computing paradigm that replicates the structure and function of the human neocortex. This paper has been superseded by the paper titled "Hierarchical Temporal Memory including HTM Cortical Learning Algorithms", above.
The HTM Learning Algorithms (PDF)
This paper documents our first generation learning algorithms, called Zeta 1. The paper has been superseded by the paper titled "Hierarchical Temporal Memory including HTM Cortical Learning Algorithms", above.
HTM: Comparison with Existing Models (PDF)
This whitepaper discusses how HTMs are different from existing models, and does a comparison of these modeling techniques. This paper will be updated in 2011. While some of the material is out of date, much of the content is still relevant.

Videos

The following presentations have been superseded by new documentation. We include them here as a reference to our previous work.

Preview of next generation HTM algorithms (video)
Jeff Hawkins presents an overview of HTM theory and then previews the second generation algorithms. March 18, 2010.
Keynote from 2008 HTM Workshop (video)
Jeff Hawkins presents Hierarchical Temporal Memory to the attendees at Numenta's 2008 HTM Workshop.
HTM: Biological Mapping to Neocortex and Thalamus (video)
This is an hour long talk describing how HTM technology maps to the anatomy and physiology or the brain. Engineers who are familiar with HTM have found this talk helpful so we decided to record it and make it available. March 20, 2007.

Academic papers

Sequence memory for prediction, inference and behaviour (PDF)
Jeff Hawkins, Dileep George and Jamie Niemasik. Philosophical Transactions of the Royal Society B 2009 364, 1203-1209. This paper describes a mechanism by which the neocortex may store sequences of patterns.
Towards a Mathematical Theory of Cortical Micro-circuits (PLoS website)
Citation: George D, Hawkins J (2009) Towards a Mathematical Theory of Cortical Micro-circuits. PLoS Comput Biol 5(10): e1000532. doi:10.1371/journal.pcbi.1000532. Authored by Dileep George and Jeff Hawkins, this paper is a peer-reviewed, scientific paper that describes in detail the mapping of our HTM theory to the neuro circuitry in the brain.
HTM Thesis: How the Brain Might Work (PDF)
Dileep George's Ph.D thesis discusses a hierarchical and temporal model for learning and recognition.

Independent sites and research papers

Links to sites and research papers from independent scientists and engineers about Hierarchical Temporal Memory and about Numenta technology.

Support for the Use of Hierarchical Temporal Memory Systems in Automated Design Evaluation: A First Experiment (PDF)
Josh Hartung, Jay McCormack & Frank Jacobus (2009). Proceedings of the ASME 2009 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference IDETC/CIE 2009 [copyright © ASME 2009].
Robust Character Recognition using a Hierarchical Bayesian Network (PDF)
Thornton, J. R., Gustafsson, T., Blumenstein, M. & Hine, T. (2006). Proceedings of the 19th Australian Joint Conference on Artificial Intelligence, AI-2006, Hobart. 1259-1264. [copyright © Springer-Verlag].
Memory-Prediction Framework for Pattern Recognition: Performance and Suitability of the Bayesian Model of Visual Cortex (PDF)
Garalevicius, Saulius. (2006). Department of Computer and Information Sciences, Temple University.
Biomimetic sensory abstraction using hierarchical quilted self-organizing maps (PDF)
Miller, J.W., Lommel, P.H. (2006) The Charles Stark Draper Laboratory, Inc.