Application of fractional theory in quantum back propagation neural network

In this paper, by applying the theory of fractional calculus to quantum back propagation (BP) neural network, a quantum BP algorithm based on the definition of fractional Grunwald-Letnikoff (G-L) is proposed. We choose the Sigmoid linear superposition function to replace the activation function of the traditional neural network to construct a fractional quantum BP neural network structure. Experimental results prove that this algorithm improves the convergence speed of the network and reduces the convergence error.