Recent Advances In Artificial Neural Networks Design And Applications

Contents......Page 0
RECENT ADVANCES IN ARTIFICIAL NEURAL NETWORKS Design and Applications......Page 1
International Series on Computational Intelligence......Page 2
Preface......Page 5
The Editors......Page 8
Contents......Page 10
A NEURO-SYMBOLIC HYBRID INTELLIGENT ARCHITECTURE WITH APPLICATIONS......Page 18
1 Introduction......Page 19
2 Knowledge Based Module for Represen-tation of Initial Domain Knowledge......Page 22
3 Extraction of Supplementary Rules via the Statistical Analysis Module......Page 23
3.1 Extraction of Correlation Rules......Page 24
3.2 Reducing the Input Dimensionality......Page 27
4 The Mapping Module......Page 28
5 The Discretization Module......Page 29
6 Refining Input Characterization......Page 31
7 Rule Extraction......Page 32
7.1 First Technique (BIO-RE)......Page 34
7.3 Third Technique (Full-RE)......Page 35
8 Rule Evaluation and Ordering Proce-dure for the Refined Expert System......Page 36
8.1 The Rule Ordering Procedure......Page 37
9 The Integrated Decision Maker......Page 38
10 Application: Controlling Water Reservoirs......Page 39
10.1 Implementation Results......Page 40
11 Application of the Statistical Approach......Page 43
12 Discussion......Page 46
References......Page 50
NEW RADIAL BASIS NEURAL NETWORKS AND THEIR APPLICATION IN A LARGE- SCALE HANDWRITTEN DIGIT RECOGNITION PROBLEM......Page 55
1 Introduction......Page 56
2 Function Approximation Models and RBF Neural Networks......Page 59
3 Reformulating Radial Basis Neural Networks......Page 62
4.1 Linear Generator Functions......Page 66
4.2 Exponential Generator Functions......Page 69
5.1 The Blind Spot......Page 71
5.2 Criteria for Selecting Generator Functions......Page 74
5.3 Evaluation of Linear and Exponential Generator Functions......Page 76
6 Learning Algorithms Based on Gradient Descent......Page 80
6.1 Batch Learning Algorithms......Page 81
6.2 Sequential Learning Algorithms......Page 83
6.3 Initialization of Supervised Learning......Page 85
7 Generator Functions and Gradient Descent Learning......Page 87
8.1 The NIST Databases......Page 89
8.2 Data Preprocessing......Page 90
8.3 Classification Tools for NIST Digits......Page 91
8.4 Role of the Prototypes in Gradient Descent Learning......Page 93
8.5 Effect of the Number of Radial Basis Functions......Page 95
8.7 Benchmarking Reformulated RBF Neural Networks......Page 99
9 Conclusions......Page 104
References......Page 106
1 Introduction......Page 111
2 Proposed Methodology......Page 114
2.1 Data Quantization Using Self-Organizing Maps......Page 115
2.2 Training Set Reduction and Classification of SOFM Prototypes for Supervised Techniques......Page 116
2.3 Fast Clustering and Labeling of SOFM Prototypes for Unsupervised Techniques......Page 117
3 Modifications of Supervised Algorithms......Page 118
3.1 Classification Using the BP Algorithm......Page 119
3.2 Classification Using the LVQ Algorithm......Page 120
3.3 The Pal-Majumder Fuzzy Classifier......Page 121
NN Algorithm......Page 122
4 Multimodular Classification......Page 123
5 Land-Cover Classification......Page 124
5.1 Supervised Classification......Page 126
5.2 Multimodular Classification......Page 129
5.3 Unsupervised Classification......Page 132
6 Summary......Page 138
References......Page 139
LEARNING FINE MOTION IN ROBOTICS: DESIGN AND EXPERIMENTS......Page 142
1 How to Find the Path?......Page 143
2 The Model-Based System......Page 144
3 The Sensor-Based System......Page 148
4 Perception Clustering......Page 150
5 Action Triggering......Page 159
6 All Together......Page 161
7 Why Use a SOM-Like Network?......Page 163
8 Planner Vs. HEKM......Page 164
9 Conclusions......Page 165
References......Page 167
1 Introduction......Page 170
2.1 Bottom-Up Propagation in a Single-Channel Block......Page 173
2.2 Class Selection......Page 174
2.3 Top-Down Propagation in a Single-Channel Block......Page 176
2.4 The Orientation System......Page 177
2.5 Class Manager......Page 178
3 Learning in MART......Page 180
3.2 Channel Credits......Page 181
3.4 Global Vigilances......Page 183
3.5 Other Characteristics......Page 185
4 Analysis of the Behavior of Certain Adaptive Parameters......Page 186
5 A Real Application Example......Page 189
6 Discussion......Page 197
References......Page 198
LATERAL PRIMING ADAPTIVE RESONANCE THEORY (LAPART)-2: INNOVATION IN ART......Page 202
1 Introduction......Page 203
2 ART-1, Stacknet, and LAPART-1......Page 205
2.2 ART-1 Architecture......Page 206
2.3 Stacknet......Page 208
2.4 LAPART-1......Page 211
3.1 Forcing Learning to Occur......Page 214
3.2 Constraints on the Input Data......Page 215
4 The Learning Theorems......Page 216
5.1 Method......Page 218
5.2 Results......Page 222
6 Discussion......Page 223
7 Conclusion......Page 224
References......Page 225
1 Introduction......Page 227
2 Agents......Page 229
3 Neural Network Role......Page 230
4 Agent Architecture......Page 231
4.1 Problem Domain......Page 232
4.2 Data......Page 233
4.3 Network Training......Page 235
5 Operation......Page 239
6 Summary......Page 241
References......Page 242
1 Introduction......Page 243
2 Process Control......Page 245
3 Use of Neural Networks in Control......Page 249
4 Case Study I: pH Control in Neutralization System......Page 255
4.1 Neutralization System......Page 256
4.2 Neural Network Control of the Neutralization System......Page 257
4.3.1 Conventional PID Controller Performance......Page 259
4.3.2 NN Controller Performance......Page 260
5 Case Study II: Adaptive Nonlinear-Model Predictive Control Using Neural Networks for Control of High Purity Industrial Distillation Column......Page 262
5.2.1 Linear Model Predictive Controller......Page 263
5.2.3 Adaptive Nonlinear-Model Predictive Controller via Neural Networks......Page 265
5.3 Identification......Page 268
5.4 Development of the Neural Network Model......Page 269
5.5 Control Application......Page 271
6.1 Binary Batch Distillation Column......Page 275
6.2 PI Controller with Neural Network as a Parameter Estimator......Page 281
6.3 Results......Page 282
7.1 Analysis of the Kettle......Page 285
7.2.2 Rule Derivation......Page 288
7.2.3 Neural Network......Page 289
7.3 Results......Page 291
8 Remarks and Future Studies......Page 295
References......Page 297
MONITORING INTERNAL COMBUSTION ENGINES BY NEURAL NETWORK BASED VIRTUAL SENSING......Page 304
1 Introduction......Page 305
2 The Engine Management System......Page 306
3 Virtual Sensor Systems......Page 308
4 Air-Fuel Ratio......Page 309
5 Combustion Monitoring Using the Spark Plug......Page 310
6 The Ignition System of a Spark-Ignition Engine......Page 312
7 Neural Networks for Use in Virtual Sensors......Page 315
8.1 The Spark Voltage Characterization Method......Page 318
8.2 Neural Network Training Procedure......Page 319
8.3.2 Unequal Sample Intervals......Page 321
8.3.4 Radial Basis Functions......Page 322
8.3.5 Discussion......Page 323
8.4 The Single-Cylinder Engine......Page 324
8.4.2 Multi-Speed Tests......Page 325
9 Conclusions......Page 327
Acknowledgment......Page 328
References......Page 329
NEURAL ARCHITECTURES OF FUZZY PETRI NETS......Page 332
1 Introduction......Page 333
2 The Generalization of the Petri Net and Its Underlying Architecture......Page 334
3 The Architecture of the Fuzzy Petri Net......Page 335
4 The Learning Procedure......Page 338
5 Interfacing Fuzzy Petri Nets with Granular Information......Page 342
6 Experiments......Page 344
7 Conclusions......Page 356
References......Page 357