Title :
Performing mathematics with neurons or neural networks trained to represent non-linear functions. II
Author_Institution :
Dept. of Comput., Univ. of Manchester Inst. of Sci. & Technol., UK
fDate :
6/24/1905 12:00:00 AM
Abstract :
For part I see IEEE WCCI-2002 World Congress Computational Intelligence. The article introduces a way of performing mathematics with neurons trained to map non-linear functions. It is novel because it enables knowledge encapsulated in trained neurons to be utilised to prescribe the weights of other neurons. The knowledge encapsulated in the trained neurons in this case is non-linear functions
Keywords :
backpropagation; mathematics computing; neural nets; nonlinear functions; symbol manipulation; mathematic calculations; neural networks; neurons; nonlinear functions; symbolic algebra; Artificial neural networks; Associative memory; Computer networks; Equations; Mathematics; Matrix decomposition; Multi-layer neural network; Neural networks; Neurons; Symmetric matrices;
Conference_Titel :
Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
0-7803-7278-6
DOI :
10.1109/IJCNN.2002.1005539