Neuro-fuzzy architectures and hybrid learning (Q5950827)

The emerging field of intelligent systems sets challenging requirements on traditional subjects of research, including fuzzy systems, neural networks, and genetic algorithms. Currently, these traditional areas are considered within the framework of soft computing, and the common goal is to increase their power by applying them jointly in the form of intelligent systems. The main idea of intelligent systems consists in combining neuro-fuzzy architectures with hybrid learning. From this point of view, the book of Professor Rutkowska presents an excellent framework for applications of soft computing methods in the area of artificial intelligence. The text can be divided into three main part. The first part (Chapters 1-3) explains the basic principles of fuzzy inference systems and neural networks. The next part (Chapters 4 and 5) deals with two main types of neuro-fuzzy architectures -- neuro-fuzzy architectures based on the Mamdani approach and neuro-fuzzy architectures based on the logical approach. The last part of the text (Chapters 6-8) is dedicated to efficient combining of the novel techniques of neural networks, fuzzy systems and genetic algorithms. The topics of hybrid learning methods and intelligent systems are here discussed also from the revolutionary perspective of perception-based computing, developed by Lofti A. Zadeh. The underlying book reviews systematically various concepts of fuzzy-neural based architectures. Their precise yet elegant analysis is based on years of both scientific and educational experience of Professor Rutkowska. Anyway, the book provides much more information, not only concentrates on the subject of her own research in implication-based neuro-fuzzy systems and hybrid learning. The covered state-of-the-art material includes rich references to literature (nearly 600 references). Numerous illustrative figures demonstrate clearly the ideas behind the discussed paradigms. Therefore, the book represents a welcome source of information for those researchers who want to get a thorough up-to-date overview in the area of intelligent systems. For the same reasons, the text might be used as a superior reference source in advanced courses on the subject, too.