Neural Field Model for Perceptual Learning

Perceptual learning is happened at the perceptual level. We combine holism and reductionism together to research on perception learning. Based on information geometry the paper will present a neural field model which is referred to understand the transformation mechanism, dynamical behavior, capability and limitation of neural network models, by the study of globally topological and geometrical structure on parameter spaces of neural networks. In this paper we will discuss neural field representation, fractal learning principle, topology approximation correction learning and dualistic correction learning algorithm. Finally the paper will give a conclusions and point out future research topics.