Abstract
This chapter presents the general background information about the Hierarchical Temporal Memory (HTM). HTM is a recently proposed cognitive learning algorithm that is intended to emulate the overall structural and functionality of the human neocortex responsible for the high-order functions such as cognition, learning and making predictions. The main properties of HTM is hierarchical structure, sparsity and modularity. HTM consists of two main parts: HTM Spatial Pooler (SP) and HTM Temporal Memory (TM). The HTM SP performs the encoding of the input data and produces sparse distributed representation (SDR) of the input pattern useful for visual data processing and classification tasks. The HTM TM detects the temporal changes in the input data and performs prediction making.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Ahmad S, Lewis M (2017) Temporal memory algorithm. Technical report, Numenta, Inc
Cui Y, Ahmad S, Hawkins J (2017) The HTM spatial Pooler- a neocortical algorithm for online sparse distributed coding. Technical report, Numenta, Inc
Fan D, Sharad M, Sengupta A, Roy K (2016) Hierarchical temporal memory based on spin-neurons and resistive memory for energy-efficient brain-inspired computing. IEEE Trans Neural Netw Learn Syst 27(9):1907–1919
Hawkins J (2004) On Intelligence. Griffin, Macmillan
Hawkins J (2016) Biological and machine intelligence. Numenta. http://numenta.com/biological-and-machine-intelligence/
Ahmad S, Hawkins J (2016) How do neurons operate on sparse distributed representations? A mathematical theory of sparsity, neurons and active dendrites. Technical report, Numenta, Inc
Hawkins J, Ahmad S (2016) Why neurons have thousands of synapses, a theory of sequence memory in neocortex. Technical report, Numenta, Inc
Hawkins J, Ahmad S, Dubinsky D (2010) Hierarchical temporal memory including HTM cortical learning algorithms. Techical report, Numenta, Inc, Palto Alto. http://www.numenta.com/htmoverview/education/HTM_CorticalLearningAlgorithms.pdf
Hawkins J, Ahmad S, Purdy S, Lavin A (2016) Biological and machine intelligence (bami). Initial online release 0.4
Ibrayev T, Krestinskaya O, James AP (2017) Design and implication of a rule based weight sparsity module in HTM spatial pooler. In: 24th IEEE international conference on electronics, circuits and systems (ICECS). IEEE, pp 274–277
Ibrayev T, Myrzakhan U, Krestinskaya O, Irmanova A, James AP (2018) On-chip face recognition system design with memristive hierarchical temporal memory. J Intell Fuzzy Syst 34(3):1393–1402
James A, Ibrayev T, Krestinskaya O, Dolzhikova I (2018) Introduction to memristive HTM circuits. In: Memristor and memristive neural networks. InTech
Krestinskaya O, James AP (2018) Feature extraction without learning in an analog spatial pooler memristive-CMOS circuit design of hierarchical temporal memory. Analog Integr Circuits Signal Process 95(3):457–465
Krestinskaya O, Dolzhikova I, James AP (2018) Hierarchical temporal memory using memristor networks: a survey. IEEE Trans Emerg Top Comput Intell 2(5):380–395. https://doi.org/10.1109/TETCI.2018.2838124
Krestinskaya O, Ibrayev T, James AP (2018) Hierarchical temporal memory features with memristor logic circuits for pattern recognition. IEEE Trans Comput-Aided Des Integr Circuits Syst 37(6):1143–1156
Krestinskaya O, James AP, Chua LO (2018) Neuro-memristive circuits for edge computing: a review. arXiv:1807.00962
Lavin A, Ahmad S (2017) Sparse distributed representations. Technical report, Numenta, Inc
Purdy S (2017) Encoding data for HTM systems. Technical report, Numenta, Inc
Spruston N (2008) Pyramidal neurons: dendritic structure and synaptic integration. Nat Rev Neurosci 9(3):206
Webster D, Popper A (1992) The mammalian auditory pathway: neuroanatomy. Springer, New York
Zyarah AM (2015) Design and analysis of a Re-configurable hierarchical temporal memory architecture. Technical report, Rochester institute of technology
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Chapter Highlights
Chapter Highlights
-
HTM is a recently proposed cognitive learning algorithm that is intended to emulate the overall structural and functionality of the human neocortex responsible for the high-order functions such as cognition, learning and making predictions.
-
The main properties of HTM is hierarchical structure, sparsity and modularity.
-
HTM consists of two main parts: HTM Spatial Pooler (SP) and HTM Temporal Memory (TM). The HTM SP performs the encoding of the input data and produces sparse distributed representation (SDR); while the HTM TM detects the temporal changes in the input data and performs prediction making.
-
HTM algorithm showed the promising results in various tasks, especially for visual data processing and classification, such as handwritten digits, face and speech recognition.
-
There are several hardware implementations of HTM; however, it is still an open problem.
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Dauletkhanuly, Y., Krestinskaya, O., James, A.P. (2020). HTM Theory. In: James, A. (eds) Deep Learning Classifiers with Memristive Networks. Modeling and Optimization in Science and Technologies, vol 14. Springer, Cham. https://doi.org/10.1007/978-3-030-14524-8_13
Download citation
DOI: https://doi.org/10.1007/978-3-030-14524-8_13
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-14522-4
Online ISBN: 978-3-030-14524-8
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)