ICANN 98

Invited Presentations.- (Back) Towards Diagrammatic Representation and Reasoning in a Connectionist Framework.- Variational Learning in Graphical Models and Neural Networks.- Synchronization: The Computational Currency of Cognition.- Applications of Vapnik’s Theory for Prediction.- Brains, Gases and Robots.- Self-Organization of Very Large Document Collections: State of the Art.- Silicon Artificial Neural Networks.- Oral Presentations: Theory: Theory I: Algorithms.- Efficient Top-Down Jacobian Evaluation of Tree-Structured Neural Networks.- Tractable Undirected Approximations for Graphical Models.- Convex Cost Functions for Support Vector Regression.- Asymptotically Optimal Choice of ?-Loss for Support Vector Machines.- Support Vector Regression with Automatic Accuracy Control.- Theory II: Dynamical systems, Time Series.- Queuing Theory for Spike Driven Synaptic Dynamics.- The Deformable Feature Map — Adaptive Plasticity for Function Approximation.- Asymptotic Behavior of a Neural Network with Dynamic External Input.- Theory III: Signal Decomposition Methods.- Independent Component Analysis for Time-Dependent Stochastic Processes.- Denoising of Sensory Data by Maximum Likelihood Estimation of Sparse Components.- Kernel PCA Pattern Reconstruction via Approximate Pre-Images.- The Pearson Mixture Model for Cluster Analysis and Data Visualisation.- Theory IV: Learning Theory.- Why Feed-Forward Networks are in a Bad Shape.- Transients and Asymptotics of Natural Gradient Learning.- An Analysis of Convergence in Generalized LVQ.- Ultrametric Structure in Autoencoder Error Surfaces.- The Dynamics of Matrix Momentum.- Dynamics of Batch Learning in Multilayer Neural Networks.- An Asymptotic Analysis of AdaBoost in the Binary Classification Case.- Least Absolute Shrinkage is Equivalent to Quadratic Penalization.- Oral Presentations: Applications: Applications I: Process Control, Diagnosis.- Neural Network Modelling of Ore Grade Spatial Variability.- Neural Virtual Sensors — Estimation of Combustion Quality in SI Engines Using the Spark Plug.- Real-Time Ion Temperature Profiles in the JET Nuclear Fusion Experiment.- Detection of Spatio-Temporal Ultrasonic Transient Families Using Filter Banks and Neural Nets.- Applications II: Image Processing.- Object Recognition with Multiple Feature Types.- Learning Gestures with Time-Delay RBF Networks.- Comparison of Adaptive Strategies for On-Line Character Recognition.- A Binary Correlation Matrix Memory k-NN Classifier.- Applications III: Speech, Telecom, Finance.- Multinet: A New Connectionist Architecture for Speech Recognition.- Simulation Support and ATM Performance Prediction.- TDNN Approach to Measuring Raindrop Sizes and Velocities.- Optimizing the Evidence.- Applications IV: Medicine, Sequence Analysis.- Adapting an Ensemble Approach for the Diagnosis of Breast Cancer.- Independent Component Analysis in Wave Decomposition of Auditory Evoked Fields.- Experiments in Gait Pattern Classification with Neural Networks of Adaptive Architecture.- Using Feature Selection to Find Inputs that Work Better as Extra Outputs.- Applications V: Hybrid Systems, General Techniques.- Analysis of Multi-Choice Questionnaires Through Self-Organizing Maps.- A Niching Genetic Algorithm for Selecting Features for Neural Network Classifiers.- The Wavelet Transform for Time Series Prediction.- Empirical Evaluation of Bayesian Sampling for Neural Classifiers.- Oral Presentations: Computational Neuroscience and Brain Theory.- State-Dependent Spatio-Temporal Restructuring of Receptive Fields in the Primary Visual Pathway.- Frequency Spectrum of Coupled Stochastic Neurons with Refractoriness.- Analysing the Context Dependence of Receptive Fields in Visual Cortex.- A Neural Model of Stereopsis.- Continuous Dynamics of Neuronal Delay Adaptation.- The Critical Synaptic Number for Rhythmogenesis and Synchronization in a Network Model of the Cerebellar Granular Layer.- A Model of Cortical Plasticity: Integration and Segregation Based on Temporal Input Patterns.- Self-Organisation and Cortical Dynamics.- Oral Presentations: Connectionist Cognitive Science and Artificial Intelligence.- Continuous Time Recurrent Neural Networks for Grammatical Induction.- Encoding Structure in Boolean Space.- Non-Compositional Representation in Connectionist Networks.- A Connectionist Model for Categorical Perception and Symbol Grounding.- Enriched Lexical Representations, Large Corpora, and the Performance of SRNs.- Oral Presentations: Autonomous Robotics and Adaptive Behavior.- Using Focus to Direct Environmental Mapping.- A Biologically Inspired Adaptive Control Architecture Based on Neural Networks for a Four-Legged Walking Machine.- 2-D Pole Balancing with Recurrent Evolutionary Networks.- GripSee: A Robot for Visually-Guided Grasping.- Dynamics of a Classical Conditioning Model.- CMAC Models Learn to Play Soccer.- Pseudo-Parametric Q-Learning Using Feedforward Neural Networks.- Continuous Q-Learning Resource Allocation Network.- Oral Presentations: Hardware and Implementations.- Sparse Distributed Memory Mapping on Partial Tree Shape Neurocomputer.- PRESENCE, a Hardware Implementation of Binary Neural Networks.- Analog VLSI Implementation of a Spike Driven Stochastic Dynamical Synapse.- Neural Dynamics in Real-Time for Large Scale Biomorphic Neural Networks.- Contents, Volume 2: Poster Presentations Poster Presentations: Theory Spotlight Presentations: Theory I: Algorithms.- A Study on Functional-Link Neural Units with Maximum Entropy Response.- Mean Field Theory Based on Belief Networks for Approximate Inference.- On the Use of Local RBF Networks to Approximate Multivalued Functions and Relations.- Learning Higher Order Boltzmann Machinces Using Linear Response.- Order Parameters for Self-Organizing Maps.- Slope Centering: Making Shortcut Weights Effective.- Spotlight Presentations: Theory II: Dynamical Systems, Time Series.- Constrained Second-Order Recurrent Networks for Finite-State Automata Induction.- Exact Learning Curves for EKF Training.- Spotlight Presentation: Theory IH: Signal Decomposition Methods.- Sparse Regression: Utilizing the Higher-Order Structure of Data for Prediction.- Poster Session I: Theory.- A Linear Programming Neural Circuit Model.- Learning Invariance Manifolds.- Developmental Evolution of an Edge Detecting Retina.- Activity Driven Update in the Neural Abstraction Pyramid.- Jacobian Neural Network Learning Algorithms.- Quadratic Concepts.- Complexity of Boolean Computations for a Spiking Neuron.- Unsupervised Time Series Segmentation by Predictive Modular Neural Networks.- ICE — An Incremental Hybrid System for Continuous Learning.- Synthesis of Probabilistic Automata in pRAM Neural Networks.- Neural Modeling of Nonlinear Differential Equations with Discrete Measurements A Lagrangian Approach.- Simple Synchrony Networks: Learning to Parse Natural Language with Temporal Synchrony Variable Binding.- Gaussian Processes for Switching Regimes.- Artificial Neural Networks as Approximators of Stochastic Processes.- Piecewise Affine Neural Networks and Nonlinear Control.- Poster Session II: Theory.- Some Complexity Results for Perceptron Networks.- Multilayer Neural Networks for Classification: A Pedagogical Theorem.- An Experimental Comparison of Neural ICA Algorithms.- The Principal Independent Components of Images.- Pattern Formation in Locally Connected Oscillatory Networks.- LOCOCODE versus PCA and ICA.- TDSEP — An Efficient Algorithm for Blind Separation Using Time Structure.- Optimal Cross-Validation Split Ratio: Experimental Investigation.- On the Convergence Properties of the Temporal Kohonen Map and the Recurrent Self-Organizing Map.- Multivariate Linear Regression on Classifier Outputs: A Capacity Study.- Poster Presentations: Applications Spotlight Presentations: Applications I: Process Control, Diagnosis.- Automated Statistical Recognition of Partial Discharges in Insulation Systems.- A Radar System with Phase-Sensitive Millimetric Wave Circuitry and Complex-Amplitude Neural Processing.- Chances and Risks of Sensor Fusion with Neural Networks: An Application Example.- Support Objects for Domain Approximation.- A Comparison of Traditional and Soft-Computing Methods in a Real-Time Control Application.- Spotlight Presentations: Applications II: Image Processing.- Efficient Local Subspace Construction for Neural Data Modeling.- Edge Detection of Multispectral Images Using the 1-D Self-Organizing Map.- Design of Cellular Neural Networks for Binary and Gray Level Image Processing.- Adaptive Illuminant Estimation Using Neural Networks.- Automatic Neural Generalized Font Identification.- Spotlight Presentations: Applications III: Speech, Telecom, Finance.- An Approach to Blin Source Separation of Speech Signals.- Data Fusion for Diagnosis in a Telecommunication Network.- Creating an Order in Distributed Digital Libraries by Integrating Independent Self-Organizing Maps.- Competitive Learning for Binary Valued Data.- Neural Network-Based Inferential Sensing.- Spotlight Presentations: Applications IV: Medicine, Sequence Analysis, Environment.- Retina Encoder Inversion for Retina Implant Simulation.- The Automated Identification of Tubercle Bacilli Using Image Processing and Neural Computing Techniques.- Black-Box Software Sensor Design for Environmental Monitoring.- Classifying Regional Seismic Signals Using TDNN-Alike Neural Networks.- A Hierarchical Self-Organizing Map Model for Sequence Recognition.- Poster Session I: Applications.- Discrete Time Backpropagation and Synaptic Delay Based Artificial Neural Networks in Chaotic Time Series Prediction.- A Learning Method of Fuzzy Inference Rules Using Vector Quantization.- Mixed Fuzzy-System and Artificial Neural Network Approach to the Automated Recognition of Mouth Expressions.- COMVIS: A Communication Framework for Computer Vision.- Implementing a Hybrid Architecture for Artificial Neural Network Applications.- Application of ANN to the Selection of a Valve from the Catalogue.- A Neural Network Approach to Functional MRI Pattern Analysis — Clustering of Time-Series by Hierarchical Vector Quantization.- Penalized Training for Serially Correlated Data.- Poster Session II: Applications.- Gaussian Mixture and Kernel Based Approaches to Blind Separation of Sources Using Neural Nets.- The Adaptive Setback Thermostat.- Architecture Optimization in Feedforward Connectionist Models.- Neural Velocity Force Control for Industrial Manipulators Contacting Rigid Surfaces.- Neural Trajectory Optimization (NTO) for Manipulator Tracking of Unknown Surfaces.- Computer Network User Behaviour Visualisation Using Self Organising Maps.- Neural Control of a Virtual Prosthesis.- Adaptive Clustering and Multidimensional Scaling of Large and High-Dimensional Data Sets.- Behavior of the Weights of a Support Vector Machine as a Function of the Regularization Parameter C.- Convergence Properties of a Modified Temporal Anti-Hebbian Model.- Volatility Prediction with Mixture Density Networks.- Introducing a Clustering Technique into Recurrent Neural Networks for Solving Large-Scale Traveling Salesman Problems.- Poster Presentations: Computational Neuroscience and Brain Theory Spotlight Presentations: Computational Neuroscience and Brain Theory (1).- Comparing Different Measures of Spatio-Temporal Patterns in Neural Activity.- Influence of Recurrent Excitation and Inhibition on Receptive Field Size and Contrast Sensitivity in Layer 4C of Macaque Striate Cortex.- Self-Organization of Shift-Invariant Receptive Fields Through Pre- and Postsynaptic Competition.- A Cortical Interpretation of ASSOMs.- Spotlight Presentations: Computational Neuroscience and Brain Theory (2).- Decoding Population Responses in Short Epochs.- Modeling Reward Dependent Activity Pattern of Caudate Neurons.- Neural Signalling: It’s a Gas.- Fast Learning of Dynamic Compensation in Motor Control.- Poster Session I: Computational Neuroscience and Brain Theory.- A One-Dimensional Frequency Map Implemented Using a Network of Integrate-and-Fire Neurons.- The Role of Spatio-Temporal Neural Response Characteristics in the Formation of Synchrony.- Three-Layered Neural Model Between Cortical Areas VI and IT.- An Interruptible Connectionist Model for Real-Time Pattern Recognition.- Implementation of Tunable Receptive Field (RF) Filters for Learning Retina Implants.- Rate and Temporal Coding with a Neural Oscillator.- Poster Session II: Computational Neuroscience and Brain Theory.- Phase Transitions in Even Cyclic Inhibitory Networks.- The Basal Ganglia Viewed as an Action Selection Device.- Parameters Estimation in the Diffusion Model for Multidimensional Neural Data.- Asynchronous Simulation of Large Networks of Spiking Neurons and Dynamical Synapses.- Novelty Learning in a Discrete Time Chaotic Network.- Spike-Based Hebbian Learning for Stimulus Discrimination.- Poster Presentations: Connectionist Cognitive Science and Artificial Intelligence Spotlight Presentation: Connectionist Cognitive Science and Artificial Intelligence.- What Type of Finite Computations Do Recurrent Neural Networks Perform.- Poster Session I: Connectionist Cognitive Science and Artificial Intelligence.- Statistical Estimation in Conceptual Spaces.- A Connectionist Account of Spanish Determiner Production.- Learning Decompositional Structures in a Network of Max-? Units with Exponents as Connection Strengths.- Fuzzy Heterogeneous Neural Networks for Signal Forecasting.- Poster Presentations: Autonomous Robotics and Adaptive Behavior Spotlight Presentations: Autonomous Robotics and Adaptive Behavior (1).- Behavioural Coordination in Acoustically Coupled Agents.- Self-Localization by Hidden Representations.- Reinforcement Learning of Collision-Free Motions for a Robot Arm with a Sensing Skin.- Multitask Pattern Recognition for Vision-Based Autonomous Robots.- Poster Session II: Autonomous Robotics and Adaptive Behavior.- Three Principles of Hierarchical Task Composition in Reinforcement Learning.- On-Line EM Algorithm and Reinforcement Learning.- Embedding Knowledge in Reinforcement Learning.- Multistage STM in a Multilayer Hebbian Learning Architecture for Local Navigation.- Diploid Robots Adapting to Fast Changing Environments.- Dynamical Adaptation of a Neural-Net Based Agent.- Poster Presentations: Hardware and Implementations Spotlight Presentations: Hardware and Implementations.- A Reconfigurable Neuroprocessor with On-Chip Pruning.- Laser Neural Network Demonstrates Data Switching Functions.- A New Stochastic Learning Algorithm for Analog Hardware Implementation.- An Analog Neural Signal Processor for Embedded Applications.- The NeuroAccess System.- Recognizing Handwritten Digits with a Dedicated Analog VLSI Feature Extractor.- Author Index.- Author Index.