zbMATH — the first resource for mathematics

Adaptive learning of polynomial networks. Genetic programming, backpropagation and Bayesian methods. (English) Zbl 1119.68158
Genetic and Evolutionary Computation. New York, NY: Springer (ISBN 0-387-31239-0/hbk). xiv, 316 p. (2006).
This book describes induction of polynomial neural networks from data. The first chapter of the book, Introduction, defines very well what this book is about: 1) neural network representation of polynomials, 2) determining a structure of the polynomial networks using evolutionary genetic programming, 3) training neural networks using backpropagation, 4) further tuning of training using probabilistic learning algorithms, and 5) model validation.
In the second chapter the authors discuss inductive genetic programming. In this chapter polynomial neural networks, especially tree-structured, are explained.
The next chapter covers tree-like polynomial neural networks in depth. The following chapter discusses fitness functions, while Chapter 5 describes how to use control micromechanisms for inductive genetic programming.
The next two chapters cover backpropagation in neural networks: first-order, second-order, and rational backpropagation, network pruning and temporal backpropagation. Chapter 8 discusses Bayesian inference techniques, i.e., the Bayesian approaches to polynomial neural network learning. Finally, Chapter 9 covers statistical validation of the discussed models.
This book may be used as a textbook for an advanced course on special topics of machine learning.

68T05 Learning and adaptive systems in artificial intelligence
68M20 Performance evaluation, queueing, and scheduling in the context of computer systems
68-01 Introductory exposition (textbooks, tutorial papers, etc.) pertaining to computer science
Full Text: DOI