DocumentCode
3285348
Title
Novel Embryonic Array with Neural Network Characteristics
Author
Samie, M. ; Dragffy, G. ; Kiely, J.
Author_Institution
IRIB, Tehran
fYear
2007
fDate
5-8 Aug. 2007
Firstpage
420
Lastpage
430
Abstract
This paper introduces a novel design approach for embryonic arrays that is able to employ all three axial characteristics of living beings: phytogeny, ontogeny, and epigenesis. Rather than using multiplexer based binary decision diagrams (BDD) it utilizes the K/N digital neurons to define the functionality of the cells. A special neural network design technique that uses Voronoi diagram (VoD) will be employed that enables neural network array-like configuration of embryonic arrays. Additionally, learning ability, geometrical extraction of genome, gate level configuration, large number of minterm and sum of products (SOP) generation for its constituent cells are some of the other advantages of the technique.
Keywords
computational geometry; field programmable gate arrays; neural nets; K/N digital neurons; Voronoi diagram; embryonic array; epigenesis; gate level configuration; geometrical genome extraction; learning ability; minterm; neural network characteristics; ontogeny; phytogeny; sum of products generation; Bioinformatics; DNA; Embryo; Euclidean distance; Field programmable gate arrays; Genomics; Hardware; Neural networks; Neurons; Phylogeny;
fLanguage
English
Publisher
ieee
Conference_Titel
Adaptive Hardware and Systems, 2007. AHS 2007. Second NASA/ESA Conference on
Conference_Location
Edinburgh
Print_ISBN
978-0-7695-2866-3
Type
conf
DOI
10.1109/AHS.2007.80
Filename
4291950
Link To Document