Title :
Dynamics of fully complex-valued neural networks
Author_Institution :
Tokyo Univ., Japan
fDate :
7/30/1992 12:00:00 AM
Abstract :
A novel neural network that processes input vectors and attractors fully in complex space using complex weights is proposed. Real and imaginary data are treated consistently with an equivalent significance in nondegenerate complex space. This network can be applied for ill-posed problems concerning realistic physical fields and continuous motion controls. The dynamics are presented and demonstrated.
Keywords :
content-addressable storage; dynamics; neural nets; Hopfield associative memory; attractor processing; complex weights; continuous motion controls; dynamics; fully complex-valued neural networks; ill-posed problems; input vector processing; nondegenerate complex space; physical field reconstruction;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19920948
