The Study Of Artificial Neural Network Approach

The current research of artificial intelligence mainly consists of the following three approaches:the symbolic approach,the connectionist approach(the artificial neural network approach)and behaviourism approach.The research object of this paper is the artificial neural network approach,which has the highest intelligence level among the three,and which the author thinks has the most development prospect and needs of the participation of the philosophy of science.Chapter one is the introduction.Chapter one firstly introduces and compares three approaches of artificial intelligence,and then elaborate the reasons of choosing the Ann approach as the resarch object through the comparision.This chapter goes on to point out that the traditional technical approach is one-sided and in which the philosophy of science is hard to take part,thus we need a definition of approach and on which the construction of the Ann approach base.Chapter two firstly introduces the concept of approach proposed by Bunge and indicates that the early Ann approach is one of the scientific approaches in bunge's approach classification,It then developed into a technical approach.Then this chapter describes the characteristics of the elements of bunge approach,provides a continuous criterion for whether the related theoretical and practical work can be considered as components of Ann approach.Chapter three to chapter five respectively construct the Ann approach in the three waves of artificial neural network,sort out their background knowledge,methodics,problematics and aims,highlights their process of development and evolution.Chapter six summarizes the characteristics and development tendency excavated during the constrution of the Ann approach and proposes some suggestions on the future development of the Ann approach:1.Strengthen the interaction between artificial neural network and neuroscience;2.The research focus of the Ann approach should shift from engineering to science.