基于神经网络的创造性计算模型的构建

Associative Hopfield Neural Networks and Establishing Model of Intuition

根据神经网络和思维科学的理论,研究了逻辑、直觉以及创造性思维的模拟方法。逻辑思维采用了神经网络的BP算法来实现,直觉思维采用了Hopfield神经网络以及交叉变异等实现方法,论述了变异联想在直觉产生中的重要作用,并且建立了模拟直觉的计算模型。采用直觉与逻辑互补构成创造性思维的观点,构建了创造性思维的认知与计算模型,给出了创造性思维模型的计算实例。

About associative Hopfield neural networks based on Hebb learning law and the existing problem,the necessary and sufficient conditions of distinguishing prototype stability were put forward and all theorems and deductions have been proved.According to the basic theory of Hopfield neural networks,the process of intuition producing is described.The associative Hopfield model sets up an associative bridge between experience and intuition.The Hebb learning law realizes the process of experience accumulating,and the models of cross and variation vividly simulate the process that the thinking of human brain mixes together and makes a sudden change,which first establish the calculating model and put forward the calculating formula for intuition producing.Finally,an example of generating fractal graphs is given to verify the model.