A Neural Network Structure Evolution Algorithm Based on e, m Projections and Model Selection Criterion
According to biological and neurophysiologic research, there is a bloom bursting of synapses in brain’s physiological growing process of newborn infants. These jillion nerve connections will be pruned and the dendrites of neurons can change their conformation in infants’ proceeding cognition process. Simulating this pruning process, a new neural network structure evolution algorithm is proposed based on e and m projections in information geometry and model selection criterion. This structure evolution process is formulated in iterative e, m projections and stopped by using model selection criterion. Experimental results prove the validation of the algorithm.
KeywordsBayesian Information Criterion Minimum Description Length Model Selection Criterion Information Geometry Flat Manifold
Unable to display preview. Download preview PDF.
- 1.Rosemary, A.R.: Cognitive Development: Psychological and Biological Perspectives. Allyn & Bacon, Boston MA (1994)Google Scholar
- 2.Kalat, J.W.: Introduction to Psychology, 4th edn. Brooks/Cole Publ. Co., Pacific Grove (1996)Google Scholar
- 3.Amari, S.: Differential-geometrical Method in Statistics. Lecture Notes in Statistics. Springer, Heidelberg (1985)Google Scholar
- 5.Liu, Y.H., Luo, S.W.: Information Geometric Analysis of Neural Network Pruning. Journal of Computer Research and Development (2006) (to be Published)Google Scholar
- 6.Akaike, H.: Information Theory and an Extension of the Maximum Likelihood Principle. In: Petrov, B.N., Csaki, F. (eds.) 2nd Int. Symposium on Info. Theory, Budapest, pp. 267–281 (1973)Google Scholar