Implementation Techniques (Neural Network Systems Techniques and Applications)

✍ Scribed by Cornelius T. Leondes

• Publisher: Academic Press
• Year: 1997
• Tongue: English
• Leaves: 421
• Edition: 1st
• Category: Library

✦ Synopsis

This volume covers practical and effective implementation techniques, including recurrent methods, Boltzmann machines, constructive learning with methods for the reduction of complexity in neural network systems, modular systems, associative memory, neural network design based on the concept of the Inductive Logic Unit, and a comprehensive treatment of implementations in the area of data classification. Numerous examples enhance the text. Practitioners, researchers,and students in engineering and computer science will find Implementation Techniques a comprehensive and powerful reference. Key Features Recurrent methods Boltzmann machines Constructive learning with methods for the reduction of complexity in neural network systems Modular systems Associative memory Neural network design based on the concept of the Inductive Logic Unit Data classification Integrated neuron model systems that function as programmable rational approximatorsWith numerous examples to enhance the text, practitioners, researchers, and students in engineering and computer science will find Implementation Techniques a uniquely comprehensive and powerful reference source

✦ Table of Contents

Front Cover......Page 1
Implementation Techniques......Page 4
Copyright Page......Page 5
Contents......Page 6
Contributors......Page 14
Preface......Page 16
I. Introduction......Page 20
II. Recurrent Neural Networks......Page 23
III. Mixed Networks......Page 36
IV. Some Open Problems......Page 65
References......Page 66
I. Introduction......Page 70
II. Relationship with Markov Chain Monte Carlo Methods......Page 72
III. Deterministic Origins of Boltzmann Machines......Page 74
IV. Hidden Units......Page 75
V. Training a Boltzmann Machine......Page 76
VI. An Example with No Hidden Unit......Page 86
VII. Examples with Hidden Units......Page 92
VIII. Variations on the Basic Boltzmann Machine......Page 100
IX. The Future Prospects for Boltzmann Machines......Page 105
References......Page 106
I. Introduction......Page 110
II. Classification......Page 114
III. Regression Problems......Page 130
IV. Constructing Modular Architectures......Page 143
V. Reducing Network Complexity......Page 148
VI. Conclusion......Page 153
VII. Appendix: Algorithms for Single-Node Learning......Page 154
References......Page 158
I. Introduction......Page 166
II. Why Modular Networks?......Page 168
IV. Input Decomposition......Page 171
V. Output Decomposition......Page 173
VI. Hierarchical Decomposition......Page 176
VII. Combining Outputs of Expert Modules......Page 178
VIII, Adaptive Modular Networks......Page 186
IX. Conclusions......Page 196
References......Page 197
I. Introduction......Page 202
II. Point Attractor Associative Memories......Page 211
III. Continuous PAAM: Competitive Associative Memories......Page 232
IV. Discrete PAAM: Asymmetric Hopfield-Type Networks......Page 251
V. Summary and Concluding Remarks References......Page 269
I. Motivation......Page 278
II. Overview......Page 281
III. Logic, Probability, and Bearing......Page 292
IV. Principle of Maximized Bearing and ILU Architecture......Page 297
V. Optimized Transmission......Page 312
VI. Optimized Transduction......Page 315
VII. ILU Computational Structure......Page 319
VIII. ILU Testing......Page 321
Appendix: Significant Marginal and Conditional ILU Distributions......Page 324
References......Page 326
I. Introduction......Page 328
II. Data Complexity......Page 330
III. Data Separability......Page 335
IV. Classifier Selection......Page 346
V. Classifier Nonlinearity......Page 360
VI. Classifier Stability......Page 373
VII. Conclusions and Discussion......Page 385
References......Page 386
I. Introduction......Page 390
II. Defining Neuronal Arithmetics......Page 393
III. Phase Space of Neuronal Arithmetics......Page 397
IV. Multimode Neuronal Arithmetic Unit......Page 406
V. Toward a Computing Neuron......Page 411
References......Page 414
Index......Page 416