DocumentCode :
3298721
Title :
A Novel Class of Generative Neural Networks
Author :
Garimella, Rama Murthy
Author_Institution :
Commun. & Networking Res. Centre, Int. Inst. of Inf. Technol., Hyderabad
Volume :
2
fYear :
2008
fDate :
18-20 Oct. 2008
Firstpage :
255
Lastpage :
259
Abstract :
In this research paper, along the lines of the goal of constructive neural networks, a novel neural network, as an alternative to Multi Layer Perceptron is proposed. Based on the idea of one step associative memories, a novel bidirectional one step associative memory is proposed. Also, an interconnected array of associative memories is proposed. Along the lines of logical basis of Hopfield neural network, two linear/non-linear dynamical systems are proposed.
Keywords :
multilayer perceptrons; Hopfield neural network; multilayer perceptron; step associative memories; Associative memory; Computer architecture; Computer networks; Error correction codes; Feedforward neural networks; Hopfield neural networks; Multilayer perceptrons; Network synthesis; Network topology; Neural networks; Neural Networks;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Natural Computation, 2008. ICNC '08. Fourth International Conference on
Conference_Location :
Jinan
Print_ISBN :
978-0-7695-3304-9
Type :
conf
DOI :
10.1109/ICNC.2008.268
Filename :
4666996
Link To Document :
بازگشت