상세정보

A representative selection of recent literature on neural networks. Intended for the uninitiated with an interest in the sciences.

정보제공 :

CONTENTS
Preface = v
Chapter 1 : Introduction = 1
 1.1 Introduction = 2
 1.2 Major Aspects of PDP Models = 9
 1.3 What Follows = 14
 1.4 Literature Overview = 16
Chapter 2 : Associative Memory = 19
 2.1 Introduction = 20
 2.2 Representation of Information by Collective States = 21
 2.3 Basic Mathematical Model = 22
 2.4 Associative Recall = 24
  2.4.1 Mathematical Modeling of Associative Recall = 25
  2.4.2 The Novelty Component Approach = 28
  2.4.3 A Stochastic Approach to Recall in Large-Scale Operations = 29
 2.5 Memory as an Adaptive Filter = 31
 2.5.1 Biological Basis for Model = 32
 2.5.2 Mathematical Model = 34
 2.6 The Sutton-Barto Model = 35
 2.7 The Heterostat = 38
  2.7.1 Mathematical Model = 41
  2.7.2 Comparison of the Heterostat with Alternative Neuronal Models = 44
 2.8 Literature Overview = 46
Chapter 3 : The perceptron = 49
 3.1 Introduction = 50
 3.2 The Structure of a Percepton = 50
 3.3 The Parallel Nature of the Perceptron's Computation Process = 54
 3.4 Basic Mathematics of Decision Surfaces = 58
  3.4.1 The Linear Machine = 59
  3.4.2 Gradient Descent Techniques = 62
 3.5 The Perceptron Covergence Theorem = 64
 3.6 Scope of the Decision Surface Methodology = 69
 3.7 Comparison ofthe perceptron with Other Learning Models = 74
 3.8 Literture Overview = 82
Chapter 4 : The Delta Rule and Learning by Back Propagation = 85
 4.1 Introduction = 86
 4.2 The Delta Rule (Widrow-Hoff Rele) = 86
  4.2.1 Change of Basis = 87
  4.2.2 Gradient Descent in the Ordinary Delta Rule = 89
  4.2.3 Extension of the Delta Rule to Statistical Learning = 90
 4.3 The Generalized Delta Rule: Learning by Back-propagation = 92
 4.4 Applications of Learning by Back-propagation = 97
  4.4.1 Bond Rating = 97
 4.5 Intractability of network learning = 97
 4.6 Literature overview = 99
Chapter 5: Some Learning Paradigms = 101
 5.1 Introduction = 102
 5.2 The Competitve Learning Paradigm = 102
 5.2.1 The place of Competitve Learning Among other learning Paradigms = 104
 5.2.2  Architectural Framework = 104
 5.3 The Linsker Model: An Example of Competitve Learning = 107
  5.3.1 Mathematical Modeling = 107
 5.3.2 Principle Component Analysis = 111
 5.3.3 Shannon Information and the Infomax Principle  = 113
 5.4 The Fukushima Models : Another Example of Competitive Learning = 116
 5.4.1 Implications of a Modified Hebbian Rule = 116
 5.4.2 The Cognitron = 117
 5.4.3 The Neocognitron = 119
 5.5 The Interactive Activation Paradigm = 120
 5.6 Comparison of the Competitive Learning and Interactive Activation Paradigms = 125
 5.7 Adaptive Resonance Theory: A Stabilized Version of Competitve Learning = 126
 5.8 Comparison of Adaptive Resonance and Resonance and Learning-by-Back-    Propagation Paradigms = 131
 5.9 Literature Overview = 134
Chapter 6 : The Hopfield and Hoppensteadt models = 135
 6.1 Introduction = 136
 6.2 The Hopfield-Tank Modle = 137
  6.2.1 Biological Background = 137
  6.2.2 Electronic Implementation = 138
  6.2.3 Discrete versus Continuous-Valued Neural Elements = 140
  6.2.4 The Minimal Energy Concept: The Motivation Behind Network Dynamics = 143
  6.2.5 Application of Hopfield Nets-The Traveling Salesman Problem = 143
  6.2.6 Application of Hopfield Nets-Problems in Vision = 148
  6.2.7 Reduction of Oscillatory Phenomena in Hopfield Nets = 150
  6.2.8 Representation of Numbers in Neural Space = 152
  6.2.9 Computational and Programming Complexity = 154
  6.2.10 Optical Implementation of Neurel Networks = 156
 6.3 The Hoppensteadt Model = 158
  6.3.1 Introduction = 158
  6.3.2 A Few Words on Neuron Physiology = 161
  6.3.3 VCON: A Voltage-Controlled Oscillator Analog of a Neuron = 162
  6.3.4 Clocks and Phase Locking = 164
 6.4 Some General Comments on Relaxation Searches = 170
 6.5 The Boltzmann Machine Learning Algorithm = 174
 6.6 Literature Overview = 176
Glossary = 177
Bibliography = 185
Index = 191