Signal processing complex components in multilayer neural network - having differentiable complex transition function associated at each network with weighted complex synapses during learning phase

摘要

During the learning phase of the signal processing, the neural network is configured according to a stochastic gradient method applied to a domain of complex signals. Real and imaginary parts with parameters (i,j,k) are taken at the ith neuron in the kth layer and the jth neuron in the (k-1)th layer. Each part of the error function is modified by a gradient term weighted by a learning speed factor. Each is adjusted by subtracting the partial differential in the error function with respect to the said part multiplied by the weighting factor. The weighting factor is calculated by increasing its value progressively until the error then increases followed by a fine adjustment. USE/ADVANTAGE - Complex signal processing, phase modulation, modulation in quadrature, radar. Non-linear processing, reduced errors.