New Theories and Methods of Neural Networks

Author: Zhang Daiyuan
Publisher:
Publish Date: 2006-11-01
Features: This book introduces the new theories and algorithms proposed by the author (Spline Weight Function Neural Network Algorithm and Algebraic Algorithm), which completely overcome the difficulties that have long troubled the academic community (such as local minima, slow convergence speed, non-convergence, and the difficulty of finding global advantages). Among them, the Spline Weight Function Neural Network Algorithm also possesses strong generalization capabilities. The book consists of 7 chapters, discussing the introduction to artificial neural networks, basic concepts, algebraic algorithms for real neural networks, global minimum analysis, algebraic algorithms for complex neural networks, Spline Weight Function Neural Networks and their learning algorithms, and statistical sensitivity analysis of neural networks. The book is logically rigorous, clearly argued, and mathematically detailed, making it a valuable reference for a wide range of scientists and engineers in related fields, as well as for doctoral and master's students in science and engineering, and senior undergraduate students. It can also serve as a training or self-study textbook for related fields.