Title :
The new weighting method of sparse neural network
Author :
Mizutani, Hikaru
Author_Institution :
Dept. of Electr. Eng., Shonan Inst. of Technol., Kanagawa, Japan
Abstract :
The author discusses a new synthesis method for a sparse neural network (NN). It is assumed that the structure of the NN is fixed. The mapping from each bit of the data to a neuron is considered, which is called a BN-mapping. There are three steps in this method. In the first step, BN-mapping is decided. In the second step, the weights are decided. And in the third step, the weights and BN-mapping are modified
Keywords :
cellular neural nets; sparse matrices; BN-mapping; sparse neural network; synthesis method; weighting method; weights; Artificial intelligence; Cellular neural networks; Equations; Error correction; Hopfield neural networks; Network synthesis; Neural networks; Neurons; Sparse matrices;
Conference_Titel :
Circuits and Systems, 1993., ISCAS '93, 1993 IEEE International Symposium on
Conference_Location :
Chicago, IL
Print_ISBN :
0-7803-1281-3
DOI :
10.1109/ISCAS.1993.394243
