The Model Of Quantum RBF Neural Network And Its Application

The quantum radial basis function neural network is a new quantum neural network model.In this model the quantum parallel computing concept is introduced to the traditional radial basis function neural network.The model combines the advantages of both quantum and radial basis function, such as computing parallelism and fast convergence speed, classification and identification of high accuracy, etc.. The model is used in infant pain expression recognition,and the results proved it is better than conventional radial basis function neural network in classification and recognition performance. The main contents are as follows:Firstly, the classic radial basis function neural network is studied, the model structure, mechanism and learning algorithm of the classic radial basis function neural network is introduced in detail. The network’s structural characteristics and network performance, such as unlimited function approximation ability and convergence and generalization ability, is analyzed.Meanwhile, several radial basis function neural network learning algorithm is distracted, and the advantages and disadvantages of each algorithm is pointed out.Secondly, the basic knowledge and related concepts of quantum computing is introduced. A new quantum radial basis function neural network is given by combining quantum computing with traditional radial basis function neural network.In this network model, the network input vector is descibed to quantum register by the quantum state conversion, so the input of the network can be simplified to an n-bit quantum superposition state, thus the network model is simplified.The data center of the network’s hidden layer is also represented to the quantum register, then in order to achieve parallel processing of the network,the adjustment of the data center can use unitary operator. After a simple binary classification problem is used to verify the accuracy of the classification and identification of the model.Finally, the quantum radial basis function neural network is applied in infant expression recognition to verify the effectiveness and practical performance of the new model in the case of higher-dimensional.