Genetic Granular Cognitive Fuzzy Neural Networks and Human Brains for Pattern Recognition
With ever-improving information technologies and high performance computational power, recent techniques in granular computing, soft computing and cognitive science have allowed an increasing understanding of normal and abnormal brain functions, especiall
- PDF / 337,979 Bytes
- 11 Pages / 430 x 660 pts Page_size
- 103 Downloads / 181 Views
2
Department of Computer Science, Wayne State University Detroit, MI 48202 USA [email protected] Department of Computer Science, Georgia State University Atlanta, GA 30302-3994 USA [email protected], [email protected] 3 Department of Psychology, Georgia State University Atlanta, GA 30303 USA [email protected], [email protected]
Abstract. With ever-improving information technologies and high performance computational power, recent techniques in granular computing, soft computing and cognitive science have allowed an increasing understanding of normal and abnormal brain functions, especially in the research of human’s pattern recognition by means of computational intelligence. It is well understood that normal brains have high intelligence to recognize different geometrical patterns, but a systematic framework of biological neural network has not yet be established. In this paper, we propose the genetic granular cognitive fuzzy neural networks (GGCFNN) in order to efficiently testify artificial neural networks’ learning capability on human’s pattern recognition in term of symmetric and similar geometry patterns. In contrast to other information systems, the GGCFNN is a highly hybrid intelligent system integrating the techniques of genetic algorithms, granular computing, and fuzzy neural networks with cognitive science for pattern recognition. Our ability to simulate biological neural networks makes it possible a more comprehensive quantitative analysis on the pattern recognition of human brains, and our preliminary experiment results would shed lights on the future research of cognitive science and brain informatics.
1
Introduction
With ever-improving information technologies and high-performance computational power, complexity and scale of brain informatics have continued to grow at an explosive pace. Recent advances in granular computing, soft computing and cognitive science have allowed an increase understanding of normal and abnormal brain functions, especially in the research of human’s pattern recognition by means of computational intelligence. It is well understood that normal N. Zhong et al. (Eds.): WImBI 2006, LNAI 4845, pp. 267–277, 2007. c Springer-Verlag Berlin Heidelberg 2007
268
C. Lin et al.
brains have high intelligence to recognize different geometrical patterns in terms of the similarity and symmetry, but a systematic framework of biological neural network has not yet been established. In addition, there is a common problem in cognitive science and computer science: How does the rule-like relational learning emerges (or why does it sometimes fail to emerge?) from the associations yielded by experience? In cognitive science’s point of view, this problem is addressed by challenging organisms with novel demands and determining how brains and behaviors are adaptively altered. In computer science, a set of training data is given to a computational model, then the problem is addressed by investigating how the networks glean useful knowledge from the given training data set and then use the discover
Data Loading...
