Title :
On modifying the weights in a modular recurrent connectionist system
Author :
Elsherif, H. ; Hambaba, M.
Author_Institution :
Intelligent Syst. Lab., Stevens Inst. of Technol., Hoboken, NJ, USA
fDate :
27 Jun-2 Jul 1994
Abstract :
A modular recurrent connectionist architecture is proposed to classify binary and continuous patterns. This system consists of three networks: one feedforward backpropagation (BP) network and two self-organization map (SOM) networks. The feedforward (basic) network is trained until a saturation error level occurs. Simultaneously, the first SOM (input control) network and the last SOM (output control) define the mapping features for the given input/output patterns. The resultant features are used by a Gaussian potential function to adjust the weights of the basic network and to classify the given patterns
Keywords :
feature extraction; feedforward neural nets; pattern classification; recurrent neural nets; self-organising feature maps; Gaussian potential function; binary pattern classification; continuous pattern classification; feedforward backpropagation network; input/output patterns; mapping features; modular recurrent connectionist system; saturation error level; self-organization map networks; weights; Computer architecture; Computer science; Feedforward systems; Hardware; Intelligent systems; Laboratories; Neurons; Performance evaluation; Psychology; Simultaneous localization and mapping;
Conference_Titel :
Neural Networks, 1994. IEEE World Congress on Computational Intelligence., 1994 IEEE International Conference on
Conference_Location :
Orlando, FL
Print_ISBN :
0-7803-1901-X
DOI :
10.1109/ICNN.1994.374221
