Title :
The cross propagation training algorithm and associated architecture
Author :
Gray, D.L. ; Michel, A.N.
Author_Institution :
Dept. of Eng., Purdue Univ. Calumet, Hammond, IN, USA
Abstract :
A novel training algorithm, called the cross propagation method (CPM), to implement binary-to-binary mappings on a four-layer perceptron-like network, is presented. The perceptron architecture is introduced, followed by further modifications to obtain a completed architecture for the cross propagation method with a four-layer structure. The associated training algorithm is called the cross propagation algorithm and is derived from some original principles of Boolean algebra followed by selected extensions. The algorithm has three modes of operations available to the user for training
Keywords :
Boolean algebra; learning systems; neural nets; Boolean algebra; binary-to-binary mappings; cross propagation training algorithm; four-layer perceptron-like network; perceptron architecture; Analog-digital conversion; Boolean algebra; Convergence; Intelligent networks; Neural networks; Neurofeedback; Neurons; Propagation losses; Quantization;
Conference_Titel :
Circuits and Systems, 1991., IEEE International Sympoisum on
Print_ISBN :
0-7803-0050-5
DOI :
10.1109/ISCAS.1991.176656