Neural networks with node gates

Function approximation problems for the ordinary neural networks may be rather difficult if the function becomes complicated, due to the necessity of big network size and the possibilities of many local minima. A promissing way to solve these difficulties a's the localization of the problem. According to this concept, a new architecture of neural network is proposed namely neural network with node gates. In this paper, a function approximation example is provided to demonstrate the better performance of the proposed networks than the ordinary neural network.