Yegnanarayana B. Artificial Neural Networks

Prentice-Hall, 2005, -479 p.Over the past fifteen years, a view has emerged that computing based on models inspired by our understanding of the structure and function of the biological neural networks may hold the key to the success of solving intelligent tasks by machines. The new field is called Artificial Neural Networks, although it is more apt to describe it as parallel and distributed processing. This introductory book is aimed at giving the basic principles of computing with models of artificial neural networks, without giving any judgment on their capabilities in solving intelligent tasks.
This text is an outgrowth of the author's teaching and research experience for the past 25 years in the areas of speech and image processing, artificial intelligence and neural networks. The principles of neural networks are closely related to such areas as pattern recognition, signal processing and artificial intelligence. Over the past 10 years many excellent books have been published in the area of artificial neural networks and many more are being published. Thus one more book like this may seem redundant. However, there seems to be still a need for a book that could be used as a text book a t an introductory level. This text is designed to meet such a demand. It must be pointed out that most of the ideas presented here have been taken from the available references and mainly from the recently published books in this area. The distinguishing feature of this book is the manner in which the various concepts are linked to provide a unified view of the subject.
The book is a self-contained, covering the fundamental principles of artificial neural networks. It can be adopted as a text book for a graduate level course. Students with basic engineering or physics or mathematics background can easily follow the topics discussed. No advanced concepts from any field are assumed. It can also be used by scientists and engineers who have an aptitude to explore new ideas in computing.
The book starts with tracing the developments in computing in general, and the trends in artificial intelligence, in particular. The prevailing notions of intelligence and intelligent tasks are discussed in the context of handling these tasks by machines. The primary reasons for the performance gaps in the current systems can be traced to the differences in the perceptions of a given input by machine and by human beings. The introductory chapter discusses the distinction between data and pattern, and between recognition and understanding, to highlight the differences in machine and human perceptions of input to a system. The chapter also deals with several pattern recognition tasks which human beings are able to perform naturally and effortlessly, whereas there are no good algorithms to implement these tasks on a machine. A brief discussion on existing models and methods of solving pattern recognition tasks is given, followed by an analysis of the need for new models of computing to deal with such tasks.Basics of Artificial Neural Networks
Activation and Synaptic Dynamics
Functional Units of ANN for Pattern Recognition Tasks
Feedforward Neural Networks
Feedback Neural Networks
Competitive Learning Neural Networks
Architectures for Complex Pattern Recognition Tasks
Applications of ANN
A - Features of Biological Neural Networks through Parallel and Distributed Processing Models
B - Mathematical Preliminaries
C - Basics of Gradient Descent Methods
D - Generalization in Neural Networks: An Overview
E - Principal Component Neural Networks: An Overview
F - Current Trends in Neural Networks