Modified self-organizing feature map algorithms for efficient digital hardware implementation

dc.contributor.author	Βασιλάς, Νικόλαος	el
dc.contributor.author	Thiran, P.	en
dc.contributor.author	Ienne, P.	en
dc.date.accessioned	2015-05-11T19:09:10Z
dc.date.issued	2015-05-11
dc.identifier.uri	http://hdl.handle.net/11400/10162
dc.rights	Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ηνωμένες Πολιτείες	*
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/us/	*
dc.source	http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=557669	en
dc.subject	Αυτο-οργανωμένοι χάρτες
dc.subject	Αλγόριθμοι
dc.subject	ψηφιακή εφαρμογή υλικού
dc.subject	διανύσματα εισόδου
dc.subject	Self-organizing maps
dc.subject	GMDH algorithms
dc.subject	digital hardware implementation
dc.subject	input vectors
dc.title	Modified self-organizing feature map algorithms for efficient digital hardware implementation	en
heal.type	journalArticle
heal.generalDescription	σε έντυπη μορφή στο γραφείο μου	el
heal.classification	Πληροφορική
heal.classification	Μηχανική υπολογιστών
heal.classification	Computer science
heal.classification	Computer engineering
heal.classificationURI	N/A-Πληροφορική
heal.classificationURI	N/A-Μηχανική υπολογιστών
heal.classificationURI	http://skos.um.es/unescothes/C00750
heal.classificationURI	http://id.loc.gov/authorities/subjects/sh85029495
heal.keywordURI	http://id.loc.gov/authorities/subjects/sh99004370
heal.keywordURI	http://id.loc.gov/authorities/subjects/sh85055435
heal.identifier.secondary	DOI: 10.1109/72.557669
heal.dateAvailable	10000-01-01
heal.language	en
heal.access	forever
heal.recordProvider	Τεχνολογικό Εκπαιδευτικό Ίδρυμα Αθήνας. Σχολή Τεχνολογικών Εφαρμογών. Τμήμα Μηχανικών Πληροφορικής Τ.Ε.	el
heal.publicationDate	1997-03
heal.bibliographicCitation	Vassilas, N., Thiran, P. and Ienne, P. (1997) Modified self-organizing feature map algorithms for efficient digital hardware implementation. IEEE Transactions on Neural Networks. [Online] 8 (2), pp.315-330. Available from: http://ieeexplore.ieee.org [Accessed 11/05/2015]	en
heal.abstract	This paper describes two variants of the Kohonen's self-organizing feature map (SOFM) algorithm. Both variants update the weights only after presentation of a group of input vectors. In contrast, in the original algorithm the weights are updated after presentation of every input vector. The main advantage of these variants is to make available a finer grain of parallelism, for implementation on machines with a very large number of processors, without compromising the desired properties of the algorithm. In this work it is proved that, for one-dimensional (1-D) maps and 1-D continuous input and weight spaces, the strictly increasing or decreasing weight configuration forms an absorbing class in both variants, exactly as in the original algorithm. Ordering of the maps and convergence to asymptotic values are also proved, again confirming the theoretical results obtained for the original algorithm. Simulations of a real-world application using two-dimensional (2-D) maps on 12-D speech data are presented to back up the theoretical results and show that the performance of one of the variants is in all respects almost as good as the original algorithm. Finally, the practical utility of the finer parallelism made available is confirmed by the description of a massively parallel hardware system that makes effective use of the best variant.	en
heal.journalName	IEEE Transactions on Neural Networks	en
heal.journalType	peer-reviewed
heal.fullTextAvailability	false