Self-Adaptive Systems for Machine Intelligence

by He, Haibo

• ISBN13: 9780470343968
• ISBN10: 0470343966
• Edition: 1st
• Format: Hardcover
• Copyright: 2011-08-09
• Publisher: Wiley-Interscience

Summary

This book will advance the understanding and application of self-adaptive intelligent systems; therefore it will potentially benefit the long-term goal of replicating certain levels of brain-like intelligence in complex and networked engineering systems. It will provide new approaches for adaptive systems within uncertain environments. This will provide an opportunity to evaluate the strengths and weaknesses of the current state-of-the-art of knowledge, give rise to new research directions, and educate future professionals in this domain.Self-adaptive intelligent systems have wide applications from military security systems to civilian daily life. In this book, different application problems, including pattern recognition, classification, image recovery, and sequence learning, will be presented to show the capability of the proposed systems in learning, memory, and prediction. Therefore, this book will also provide potential new solutions to many real-world applications.

Author Biography

Haibo He, PhD, is Assistant Professor in the Department of Electrical, Computer, and Biomedical Engineering at the University of Rhode Island. His primary research interest is computational intelligence and self-adaptive systems, including optimization and prediction, biologically inspired machine intelligence, machine learning and data mining, hardware design (VLSI/FPGA) for machine intelligence, as well as various application fields such as smart grid, sensor networks, and cognitive radio networks.

Table of Contents

Preface	p. xi
Acknowledgments	p. xv
Introduction	p. 1
The Machine Intelligence Research	p. 1
The Two-Fold Objective: Data-Driven and Biologically Inspired Approaches	p. 4
How to Read This Book	p. 8
Part I: Data-Driven Approaches for Machine Intelligence (Chapters 2, 3, and 4)	p. 8
Part II: Biologically-Inspired Approaches for Machine Intelligence (Chapters 4, 5, and 6)	p. 9
Summary and Further Readings	p. 10
References	p. 10
Incremental Learning	p. 13
Introduction	p. 13
Problem Foundation	p. 13
An Adaptive Incremental Learning Framework	p. 14
Design of the Mapping Function	p. 19
Mapping Function Based on Euclidean Distance	p. 19
Mapping Function Based on Regression Learning Model	p. 20
Mapping Function Based on Online Value System	p. 23
A Three-Curve Fitting (TCF) Technique	p. 23
System-Level Architecture for Online Value Estimation	p. 26
Case Study	p. 29
Incremental Learning from Video Stream	p. 30
Feature Representation	p. 30
Experimental Results	p. 31
Concept Drifting Issue in Incremental Learning	p. 33
Incremental Learning for Spam E-mail Classification	p. 37
Data Set Characteristic and System Configuration	p. 37
Simulation Results	p. 38
Summary	p. 39
References	p. 41
Imbalanced Learning	p. 44
Introduction	p. 44
The Nature of Imbalanced Learning	p. 44
Solutions for Imbalanced Learning	p. 48
Sampling Methods for Imbalanced Learning	p. 49
Random Oversampling and Undersampling	p. 50
Informed Undersampling	p. 51
Synthetic Sampling with Data Generation	p. 52
Adaptive Synthetic Sampling	p. 53
Sampling with Data Cleaning Techniques	p. 56
Cluster-Based Sampling Method	p. 57
Integration of Sampling and Boosting	p. 59
Cost-Sensitive Methods for Imbalanced Learning	p. 62
Cost-Sensitive Learning Framework	p. 62
Cost-Sensitive Data Space Weighting with Adaptive Boosting	p. 63
Cost-Sensitive Decision Trees	p. 65
Cost-Sensitive Neural Networks	p. 66
Kernel-Based Methods for Imbalanced Learning	p. 68
Kernel-Based Learning Framework	p. 68
Integration of Kernel Methods with Sampling Methods	p. 69
Kernel Modification Methods for Imbalanced Learning	p. 70
Active Learning Methods for Imbalanced Learning	p. 71
Additional Methods for Imbalanced Learning	p. 73
Assessment Metrics for Imbalanced Learning	p. 75
Singular Assessment Metrics	p. 75
Receiver Operating Characteristics (ROC) Curves	p. 77
Precision-Recall (PR) Curves	p. 79
Cost Curves	p. 80
Assessment Metrics for Multiclass Imbalanced Learning	p. 80
Opportunities and Challenges	p. 82
Case Study	p. 84
Nonlinear Normalization	p. 84
Data Sets Distribution	p. 88
Simulation Results and Discussions	p. 92
Summary	p. 98
References	p. 100
Ensemble Learning	p. 108
Introduction	p. 108
Hypothesis Diversity	p. 108
Q-Statistics	p. 109
Correlation Coefficient	p. 110
Disagreement Measure	p. 110
Double-Fault Measure	p. 110
Entropy Measure	p. 111
Kohavi-Wolpert Variance	p. 111
Interrater Agreement	p. 111
Measure of Difficulty	p. 112
Generalized Diversity	p. 112
Developing Multiple Hypotheses	p. 114
Bootstrap Aggregating	p. 114
Adaptive Boosting	p. 114
Subspace Learning	p. 119
Stacked Generalization	p. 120
Mixture of Experts	p. 122
Integrating Multiple Hypotheses	p. 123
Case Study	p. 126
Data Sets and Experiment Configuration	p. 127
Simulation Results	p. 128
Margin Analysis	p. 129
A Short History of Margin Analysis	p. 129
Margin Analysis for Ensemble Learning	p. 131
Summary	p. 136
References	p. 137
Adaptive Dynamic Programming for Machine Intelligence	p. 140
Introduction	p. 140
Fundamental Objectives: Optimization and Prediction	p. 141
ADP for Machine Intelligence	p. 143
Hierarchical Architecture in ADP Design	p. 143
Learning and Adaptation in ADP	p. 146
The Action Network	p. 148
The Reference Network	p. 150
The Critic Network	p. 152
Learning Strategies: Sequential Learning and Cooperative Learning	p. 154
Case Study	p. 155
Summary	p. 160
References	p. 161
Associative Learning	p. 165
Introduction	p. 165
Associative Learning Mechanism	p. 165
Structure Individual Processing Elements	p. 166
Self-Determination of the Function Value	p. 167
Signal Strength for Associative Learning	p. 168
The Associative Learning Principle	p. 169
Associative Learning in Hierarchical Neural Networks	p. 173
Network Structure	p. 173
Network Operation	p. 174
Feedforward Operation	p. 174
Feedback Operation	p. 177
Case Study	p. 180
Hetero-Associative Application	p. 180
Auto-Associative Application	p. 182
Panda Image Recovery	p. 183
Chinese Character Recognition and Recovery	p. 184
Associative Memory for Online Incremental Learning	p. 186
Summary	p. 187
References	p. 188
Sequence Learning	p. 190
Introduction	p. 190
Foundations for Sequence Learning	p. 190
Sequence Learning in Hierarchical Neural Structure	p. 194
Level 0: A Modified Hebbian Learning Architecture	p. 195
Level 1 to Level N: Sequence Storage, Prediction, and Retrieval	p. 198
Sequence Storage	p. 198
Sequence Prediction	p. 201
Prediction Mechanism	p. 201
Activation of Prediction Neuron	p. 205
Time-Controlled Multiplexer	p. 205
Sequence Retrieval	p. 207
Memory Requirement	p. 207
Learning and Anticipation of Multiple Sequences	p. 208
Case Study	p. 211
Summary	p. 212
References	p. 213
Hardware Design for Machine Intelligence	p. 217
A Final Comment	p. 217
References	p. 220
List of Abbreviations	p. 222
Index	p. 227
Table of Contents provided by Ingram. All Rights Reserved.