國立高雄大學圖資館 |

Language: English

Back

A Mathematical Formalization of Hier...

Mnatzaganian, James W.

A Mathematical Formalization of Hierarchical Temporal Memory's Spatial Pooler for use in Machine Learning.

Record Type:	Electronic resources : Monograph/item
Title/Author:	A Mathematical Formalization of Hierarchical Temporal Memory's Spatial Pooler for use in Machine Learning.
Author:	Mnatzaganian, James W.
Published:	Ann Arbor : ProQuest Dissertations & Theses, 2016
Description:	109 p.
Notes:	Source: Masters Abstracts International, Volume: 55-05.
Notes:	Adviser: Dhireesha Kudithipudi.
Contained By:	Masters Abstracts International55-05(E).
Subject:	Computer engineering.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10117687
ISBN:	9781339790619

A Mathematical Formalization of Hierarchical Temporal Memory's Spatial Pooler for use in Machine Learning.
Mnatzaganian, James W.

A Mathematical Formalization of Hierarchical Temporal Memory's Spatial Pooler for use in Machine Learning. - Ann Arbor : ProQuest Dissertations & Theses, 2016 - 109 p.

Source: Masters Abstracts International, Volume: 55-05.

Thesis (M.S.)--Rochester Institute of Technology, 2016.

This item is not available from ProQuest Dissertations & Theses.

Hierarchical temporal memory (HTM) is an emerging machine learning algorithm, with the potential to provide a means to perform predictions on spatiotemporal data. The algorithm, inspired by the neocortex, consists of two primary components, namely the spatial pooler (SP) and the temporal memory (TM). The SP is utilized to map similar inputs into generalized sparse distributed representations (SDRs). Those SDRs are then utilized by the TM, which performs sequence learning and prediction. One challenge with HTM is ensuring that proper SDRs are generated from the SP. If the SDRs are not generalizable, the TM will not be able to make proper predictions.

ISBN: 9781339790619Subjects--Topical Terms:

212944
Computer engineering.

A Mathematical Formalization of Hierarchical Temporal Memory's Spatial Pooler for use in Machine Learning.
LDR:02857nmm a2200325 4500 001 502064
005 20170619070722.5
008 170818s2016 ||||||||||||||||| ||eng d
020 $a 9781339790619
035 $a (MiAaPQ)AAI10117687
035 $a AAI10117687
040 $a MiAaPQ $c MiAaPQ
100 1 $a Mnatzaganian, James W. $3 766036
245 1 2 $a A Mathematical Formalization of Hierarchical Temporal Memory's Spatial Pooler for use in Machine Learning.
260 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2016
300 $a 109 p.
500 $a Source: Masters Abstracts International, Volume: 55-05.
500 $a Adviser: Dhireesha Kudithipudi.
502 $a Thesis (M.S.)--Rochester Institute of Technology, 2016.
506 $a This item is not available from ProQuest Dissertations & Theses.
520 $a Hierarchical temporal memory (HTM) is an emerging machine learning algorithm, with the potential to provide a means to perform predictions on spatiotemporal data. The algorithm, inspired by the neocortex, consists of two primary components, namely the spatial pooler (SP) and the temporal memory (TM). The SP is utilized to map similar inputs into generalized sparse distributed representations (SDRs). Those SDRs are then utilized by the TM, which performs sequence learning and prediction. One challenge with HTM is ensuring that proper SDRs are generated from the SP. If the SDRs are not generalizable, the TM will not be able to make proper predictions.
520 $a This work focuses on the SP and its corresponding output SDRs. A single unifying mathematical framework was created for the SP. The primary learning mechanism was explored, where a maximum likelihood estimator for determining the degree of permanence update was proposed. The boosting mechanisms were studied and found to only be relevant during the initial few iterations of the network. Observations were made relating HTM to well-known algorithms such as competitive learning and attribute bagging. Methods were provided for using the SP for classification as well as dimensionality reduction. Empirical evidence verified that given the proper parameterizations, the SP may be used for feature learning.
520 $a Similarity metrics were created for scoring the SDRs produced by the SP. The overlap metric proved that the SP is extremely robust to noise. The SP was able to produce similar outputs for a given input, provided the noise did not cause the input to change classes. This overlap metric was further utilized to create a classifier for novelty detection. The SP proved to be able to withstand more noise than the well-known support vector machine (SVM).
590 $a School code: 0465.
650 4 $a Computer engineering. $3 212944
650 4 $a Statistics. $3 182057
690 $a 0464
690 $a 0463
710 2 $a Rochester Institute of Technology. $b Computer Engineering. $3 766037
773 0 $t Masters Abstracts International $g 55-05(E).
790 $a 0465
791 $a M.S.
792 $a 2016
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=10117687