Machine Intelligence: Quo Vadis? -- CONTENTS -- Forewords -- Preface -- Chapter 1 INTRODUCTION -- Quo Vadis, Computational Intelligence? -- 1 Introduction. -- 2 The ultimate goals of CI. -- 3 A roadmap to creative systems. -- 3.1 Threshold neurons and perceptrons -- 3.2 Increasing complexity of internal PE states -- 3.3 Increasing complexity of PE interactions -- 3.4 Beyond the vector space concept -- 3.5 Flexible incremental approaches -- 3.6 Evolution of networks -- 3.7 Transition to symbolic processing -- 3.8 Up to the brains and the societies of brains -- 4 Problems pointed out by experts -- 4.1 General CI problems related to human-like intelligence -- 4.2 General problems within certain CI disciplines -- 5 Conclusions -- Acknowledgments -- References -- Chapter 2 MATHEMATICAL TOOLS FOR MACHINE INTELLIGENCE -- Mappings between High-dimensional Representations in Connectionistic Systems -- 1 Introduction -- 2 Fixed and variable-basis approximation -- 3 Upper bounds on rates of approximation of multivariable functions -- 4 Rates of approximation of real-valued Boolean functions -- 5 Upper bounds on variation -- 6 Lower bounds on variation -- 7 Discussion -- Acknowledgement -- References -- The Stimulating Role of Fuzzy Set Theory in Mathematics and its Applications -- 1 Introduction -- 2 The stage of straightforward fuzzification -- 3 The stage of explosion of possible fuzzifications -- 4 The current stage -- 4.1 Standardization, axiomatization and L-fuzzification -- 4.2 Applications of fuzzy set theory -- References -- K-order Additive Fuzzy Measures: A New Tool for Intelligent Computing -- 1 Introduction -- 2 k-order additive belief functions -- 3 Construction and identification of 2-order additive belief functions -- 4 Conclusions -- Acknowledgement -- References.

On-line Adaptation of Recurrent Radial Basis Function Networks using the Extended Kalman Filter -- 1 Introduction -- 2 Dynamic system modeling using the RRBF networks -- 3 State and parameter estimation applying EKF -- Time update equations -- 4 On-line adaptation of the RRBF network structure -- 4.1 RRBF network growing -- 4.2 RRBF network pruning -- Parameter saliency -- Parameter significance -- Time varying parameter significance -- State estimation error update -- Neuron pruning -- 5 Experiments -- 5.1 Non-stationary Mackey-Glass time series prediction -- 5.2 Non-stationary Lorenz time series prediction -- 5.3 Resolving noise/non-stationarity dilemma -- 6 Conclusions -- 7 References -- Iterative Evaluation of Anytime PSGS Fuzzy Systems -- 1 Introduction -- 2 Anytime systems -- 2.1 Iterative algorithms -- 2.2 Anytime systems with modular architecture -- 2.3 Fuzzy tools in anytime environment -- 3 SVD-based complexity reduction -- 3.1 The Basic SVD algorithm -- 4 Iterative evaluation of PSGS fuzzy systems -- 4.1 Transformation -- 4.2 Error estimation -- 5 Illustrative example -- 6 Conclusions -- Acknowledgements -- References -- Kolmogorov's Spline Network -- 1 Introduction -- 2 The Basic Theorem -- 3 The Ensemble Approach (EA) -- 4 Conclusions -- References -- Extended Kalman Filter Based Adaptation of Time-varying Recurrent Radial Basis Function Networks Structure -- 1 Introduction -- 2 RRBF network as the NARX model -- 3 State and parameter estimation applying EKF -- 4 On-line adaptation of RRBF network structure -- 4.1 RRBF network growing -- 4.2 RRBF network pruning -- 5 Experiments -- 5.1 Non-stationary Mackey-Glass time series prediction -- 5.2 Non-stationary Lorenz time series prediction -- 6 Conclusions -- 7 References -- A Multi-NF Approach with a Hybrid Learning Algorithm for Classification -- 1 Introduction.

2 Neuro-fuzzy systems for classification -- 3 Multi-NF approach -- 4 Rule generation algorithms -- 5 Hybrid learning methods -- 6 Classification examples -- 7 Conclusions -- References -- A Neural Fuzzy Classifier Based on MF-ARTMAP -- 1 Introduction -- 2 Motivation of the project -- 3 Description of the Method -- 3.1 Description of the neural network topology -- 3.2 Parallel MF-ARTMAP -- 4 Experimental results -- 4.1 Accuracy assessment -- 4.2 Experiments on benchmark data -- 4.3 Experiments on real-world data -- 4.3.1 Experiments on multi-spectral image data -- 4.3.2 Experiments on financial fraud data -- 5 Conclusion -- References -- Mathematical Properties of Various Fuzzy Flip-flops as a Basis of Fuzzy Memory Modules -- 1 Introduction -- 2 JK Fuzzy Flip-Flop -- 2.1 Max-Min operation system -- 2.2 Alegebraic operation system -- 2.3 Bounded operation system -- 2.4 Drastic operation system -- 3 D and T fuzzy flip-flops -- 3.1 D fuzzy flip-flop -- 3.2 T fuzzy flip-flop -- 3.3 SR fuzzy flip-flop -- 4 Performance of fuzzy flip-flops -- 5 Summary -- References -- Generalized T-Operators -- 1 Introduction -- 2 T-operators, negation and some basic properties -- 3 Uninorms -- 4 Nullnorms -- 5 Compensative operations -- 6 Averaging operators -- 7 Absorbing-norms -- 8 distance-based evolutionary operators -- 9 The structure of evolutionary operators -- 10 Properties of distance-based operators -- 11 Distance-based operators as parametric evolutionary operators -- 12 Entropy-based fuzzy connectives -- 13 Some other families of generalized operations -- Acknowledgement -- References -- Fuzzy Rule Extraction from Input/Output Data -- 1 Introduction -- 2 Fuzzy systems -- 3 The Levenberg-Marquardt algorithm -- 4 The bacterial algorithm -- 4.1 The encoding method: -- 4.2 The bacterial evolutionary algorithm: -- 4.2.1 Generating the initial population.

4.2.2 Bacterial mutation -- 4.3 Gene transfer -- 4.4 Stop condition -- 5 Clustering-Based Rule Extraction Technique -- 6 Fuzzy C-Means Clustering -- 7 Projection-Based Rule Extraction Technique -- 8 Trapezoidal Approximation Technique -- 9 Merging Scheme -- 10 PROCEDURE find_MD_cluster -- 11 Conclusions -- References -- Knowledge Discovery from Continuous Data Using Artificial Neural Networks -- 1 Introduction -- 2 Network training and pruning algorithm -- 3 Approximating Hidden Unit Activation Function -- 4 Rule Generation -- 5 Illustrative Examples -- 6 Conclusion -- References -- Chapter 3 ADVANCED APPLICATIONS WITH MACHINE INTELLIGENCE -- Review of Fuzzy Logic in the Geological Sciences: Where We Have Been and Where We Are Going -- 1 INTRODUCTION -- 2 REVIEW OF AREAS OF GEOLOGY WHERE FUZZY LOGIC HAS BEEN EMPLOYED: (Where We Have Been) -- 3 GEOTECHNICAL ENGINEERING -- 4 SURFACE HYDROLOGY -- 5 Subsurface Hydrology -- 6 Ground Water Risk Assessment -- 7 Hydrocarbon Exploration -- 8 Earthquake Seismology -- 9 Soil science and landscape development -- 10 Deposition of sediments -- 11 Miscellaneous -- 12 Conclusions -- Acknowledgements -- References -- Bayesian Neural Networks in Prediction of Geomagnetic Storms -- 1 Introduction -- 2 Bayesian probability theory -- 3 Neural Networks as probabilistic models -- 4 Starting points to the application -- 5 Results of GMS predictions -- References -- Adaptation in Intelligent Systems: Case Studies from Process Industry -- 1 Introduction -- 2 Smart Adaptive Systems -- 3 Software Sensors -- 3.1 Basic Principles -- 3.2 Case - Kappa Number Prediction in Pulp Cooking -- 3.3 Case - Adaptive Modelling of Carbon Dioxide in a Burning Process -- 4 Adaptive Intelligent Control -- 4.1 Basic Principles -- 4.2 Case - Rotary Dryer Control -- 5 Diagnostics -- 5.1 Basic Principles.

5.2 Nozzle Clogging in Continuous Casting of Steel -- 5.3 Web Break Indicator for a Paper Machine -- 6 Quality Control -- 6.1 Basic Principles -- 6.2 Case - Quality Control of a TMP Plant -- 7 Conclusions -- References -- Estimation and Control of Non-linear Process Using Neural Networks -- 1 Introduction -- 2 Non - linear System Identification -- 3 Non-linear Control -- 3.1 Simulation Results of Non-linear Control Using Parameter Estimation -- 4 Conclusions -- References -- The Use of Non-linear Partial Least Square Methods for On-line Process Monitoring as an Alternative to Artificial Neural Networks -- 1 Introduction -- 2 The Method -- 2.1 The PLS Method -- 2.2 The NNPLS Approach -- 2.3 The PEANO System -- 2.4 Tests using ANN and NLPLS inside PEANO -- 3 Conclusions -- References -- Recurrent Neural Networks for Real-time Computation of Inverse Kinematics of Redundant Manipulators -- 1 Introduction -- 2 Problem formulation -- 3 Recurrent Neural network models -- 3.1 The Lagrangian neural network -- 3.2 The primal-dual neural network -- 3.3 The dual neural network -- 4 Simulation results -- 4.1 Simulation results of the Lagrangian network and the prima-dual neural network -- 4.2 Simulation results of the dual neural network -- 5 Concluding remarks -- References -- Towards Perception-based Fuzzy Modeling: An Extended Multistage Fuzzy Control Model and its Use in Sustainable Regional Development Planning -- 1 Introduction -- 2 Extending Bellman and Zadeh's approach to decision making and control under fuzziness -- 3 Extending multistage decision making (control) in Bellman and Zadeh's setting -- 4 Socioeconomic regional development planning under fuzziness -- 4.1 A multistage fuzzy decision making model of regional develoment planning -- 5 Concluding remarks -- Bibliography -- Chapter 4 MACHINE INTELLIGENCE FOR HIGH LEVEL INTELLIGENT SYSTEMS.

Neural Network Models for Vision.