Abstract
Unsupervised learning plays an important role in the neural networks. Focusing on the unsupervised mechanism of neural networks, a novel generalized goodness criterion for the unsupervised neural learning of visual perception based on the martingale measure is proposed in the paper. The differential geometrical structure is used as the framework of the whole inference and spatial statistical description with adaptive attribute is embedded in the corresponding nonlinear functional space. Consequently the integration of optimization process and computational simulation with the Neo-Darwinian paradigm is obtained. And the generalization of the guidance for the evolutionary learning in the neural net framework, the convergence of the goodness and process of the evolution guaranteed by the mathematical features are discussed. This criterion has generic significance in the field of machine vision and visual pattern classification.
Similar content being viewed by others
References
Erkki Oja, Jouko Lampinen. Unsupervised learning for feature extraction, Computational intelligence—imitating life. In: Jacek M Zuradaet al, eds. New York: IEEE Press, 1994. 13–22
Liu Jiangin. A neural classifier model based on unsupervised martingale measure. Application Research of Computers (in Chinese), 1996, 13(Suppl. 1): 19–22
Zhao Dagang, Zhu Yingshan. Applied random process (in Chinese). Beijing: Mechanical Industry Press, 1990. 213
Author information
Authors and Affiliations
Additional information
Synopsis of the author Liu Jianqin, associate professor, born in Nov. 1, 1964. His research fields focus on artificial life, evolutionary computation, chaotic dynamics and nonlinear pattern recognition. He has published more than 50 papers. From Sept. 9, 1994 to Sept. 1995, as a visiting scholar, he worked in Information and Communication R & D Center, Ricoh Co. Ltd. Yokohama, Japan.
Rights and permissions
About this article
Cite this article
Liu, J. A generalized goodness criterion for unsupervised neural learning of visual perception. J. Cent. South Univ. Technol. 3, 166–170 (1996). https://doi.org/10.1007/BF02652198
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF02652198