Gielen S., Kappen B. (eds.) Artificial Neural Networks, ICANN 1993

Springer, 1993. — 1117 p.Proceedings of the International Conference on Artificial Neural Networks, Amsterdam, The Netherlands, 13-16 September 1993.
This book contains the proceedings of the International Conference on Artificial Neural Networks which was held between September 13 and 16 in Amsterdam. It is the third in a series which started two years ago in Helsinki and which last year took place in Brighton. Thanks to the European Neural Network Society, ICANN has emerged as the leading conference on neural networks in Europe.
Neural networks is a field of research which has enjoyed a rapid expansion and great popularity in both the academic and industrial research communities. The field is motivated by the commonly held belief that applications in the fields of artificial intelligence and robotics will benefit from a good understanding of the neural information processing properties that underlie human intelligence. Essential aspects of neural information processing are highly parallel execution of computation, integration of memory and process, and robustness against fluctuations. It is believed that intelligent skills, such as perception, motion and cognition, can be easier realized in neuro-computers than in a conventional computing paradigm.
This requires active research in neurobiology to extract computational principles from experimental neurobiological findings, in physics and mathematics to study the relation between architecture and function in neural networks, and in cognitive science to study higher brain functions, such as language and reasoning.
Neural networks technology has already lead to practical methods that solve real problems in a wide area of industrial applications. The clusters on robotics and applications contain sessions on various sub-topics in these fields.
These proceedings contain an up-to-date overview of research and applications, at universities as well as in industry. The large number of contributions have passed through a rigorous process of reviewing. Our program committee has been invaluable in this process.Plenary Contributions
Dynamic coupling in cortical neural networks
Keeping neural networks simple
Principles from Neurobiology
Memory and selforganization – oral contributions
The autoassociative hypothesis places constraints on hippocampal organization
Metastability of network attractor and dream sleep
Somatosensory cortical maps: reorganization following postontogenetic plasticity-experiments and theory
Adequate input for learning in attractor neural networks
Memory and selforganization – poster contributions
Neurobiological modelling and structured neural networks
Model analysis of associative learning in the photoreceptor of marine mollusc
A neural network model for motor shapes learning and programming
Learning through adaptive value: a model working in a variable environment
Improving categorization with CALM maps
A simple self-organizing neural network architecture for selective visual attention
Detection of coincidences and generation of hypotheses – a proposal for an elementary cortical function
DIVA: a self-organizing neural network model for motor equivalent speech production
Adaptive non-uniform AID conversion achieved with an unsupervised learning rule maximizing information-theoretic entropy
Optimal topology-preservation using self-organising logical neural networks
Incorporation of neurobiological aspects of Aplysia's associative conditioning in neural networks for on-line pattern detection
Description on the use of the autogenerative nodal memory model (ANM) as controlling element of an autonomously responsive system
Human memory-neurocomputer (MeNeCo project): structure for reverbation of the information in N-peaked nets (in STMemory)
Visuo-motor interaction – oral contributions
Neural representation of saccadic eye movements in monkey superior colliculus
A self-organizing neural network for learning a body-centered invariant representation of 3-D target position
Dynamic field approach to target selection in gaze control
Differences in synaptic input and excitability between superficial and deep pyramidal cells in the cat sensorimotor cortex
Visuo-motor interaction – poster contributions
An adaptive sensory fusion approach for the superior colliculus
A neural network model for spatial information representation
A dynamical model for the generation of curved trajectories
Functional organisation in the cerebellum
Activation and contraction of a muscle
The visual system – oral contributions
Correlated neuronal activity and behaviour
Map structure from pinwheel position
Emergence of transient oscillations in an ensemble of neurons
A distributed multicolumnar system for primary cortical analysis of real-world scenes
The visual system – poster contributions
Singularities in cortical orientation and direction maps: vortices, strings and bubbles
A new model for spatial frequency and orientation tuning in the visual cortex based on delayed inputs from the retina
Cascaded intra cortical inhibition: modeling connection schemes on a large scale simulator
Hidden assembly dynamics and correlated neuronal responses
A model for latencies in the visual system
A neural architecture for textured color image segmentation and recognition
Dynamics of single neurons – oral contributions
Activity-dependent modification of intrinsic neuronal properties
Implications of and activity-dependent neurite outgrowth for developing neural networks
PCA properties of interneurons
Temporal distributed processing-TDP: a time-based processing scheme accounts for time dependent receptive fields and representational maps
Dynamics of single neurons – poster contributions
Stochastic specificity in neural interaction
A computer simulation model of backwards feedback across synapse via arachidonic acid
Study of a self-learning artificial neuron model
Simulation study on calcium-activated dynamics of compartment dendrite model
On the adaptive capabilities of pulse-coded cable neurons
A local approximation of the cable equation for implementing a local interaction model
Effect of glutamate uptake on the response dynamics of the retinal horizontal cell
Robotics
Robot vision – oral contributions
Neural networks for robot eye-hand coordination
Unsupervised formation of feature detectors using residual inputs
Geometry-driven diffusion: coupled diffusion maps as a model for excitatory and inhibitory behaviour in vision
SPIN: learning and forgetting surface classifications with dynamic neural networks
Robot vision – poster contributions
Motion parallax from catastrophes in scale-space
Stability and convergence control in cooperative integration networks
Towards a neural architecture for unified visual contrast and brightness perception
Fuzzy Kohonen clustering networks for reducing search space in 3-D object recognition
An active resistor mesh embedding cortical visual processing
A fast BCSIFCS algorithm for image segmentation
Robot control – oral contributions
Neural architecture for robot planning
From situations to actions: motion behavior learning by selforganization
Application of Q-learning in robot grasping tasks
I/O-stability for robot control with a global neural net inverse model in the feedback loop
Robot control – poster contributions
A self-organizing neural network for robot motion planning
Evolved recurrent dynamical networks use noise
The Bellmann Mapping Machine for nonlinear approximation in control policy space
A real-time robot demonstration controlled by the BSP400 neurocomputer
First results on stable adaptive robot control with RBF networks
Fuzzy inference, radial basis functions, and control of flexible robotic manipulators
A recurrent trajectory storage network with parceling of the workspace
Node allocation and topographical encoding NATEnet for inverse kinematics of a 6-DOF robot arm
Learning optimal control using neural networks
A Boolean net as an adaptive and universal robot control
Transforming occupancy grids under robot motion
Complex tasks and robots
Cognitive Connectionism
Neural networks, natural language and artificial intelligence – oral contributions
NN approaches to natural language: context and trends
Integration of ANNs and dynamic concepts to an adaptive and self-organizing agent
Learning fuzzy production rules for approximate reasoning in connectionist production systems
A representational architecture for nonmonotonic inheritance structures
Neural networks, natural language and artificial intelligence – poster contributions
Spectral timing and integration of multimodel systemic processes
Net-to-rule transformation using penalty functions
Teaching homing behaviour to a neural state machine
Neural/iconic understanding of the visual world
Connectionist "symbol" systems: cognition as the sum of analogy, exemplar manipulation and language
Symbol-manipulation with attractor neural networks
A consideration on visual strategy of fovea and saccadic movement from experimental results
Iconic language representation in a recursive neural system
Miniature language acquisition tasks using dynamic weightless systems
Activity curvature: a new approach to perception
An outline for a theory of the emotions
Natural language and speech recognition – oral contributions
Alpha-Beta TDNN implement "fuzzy" connectionist time alignment in speech recognition
Continuous speech recognition predictive systems
Handling context-dependencies in speech by LVQ
Natural language and speech recognition – poster contributions
An analytically transparent network for sequence recognition
Conceptual clustering using a connectionist approach
An extended Kohonen feature map for sentence recognition
Neural nets that discuss: a general model of communication based on self-organizing maps
Neural network and nearest neighbor comparison of speaker normalization methods for vowel recognition
Speech recognition by hierarchical segment classification
Visualization and classification of voice quality with the selforganizing map
Weighted distance measure for speaker-independent digit recognition with hidden-control neural network
Modulation-frequency encoding of speech with applications to neural speech recognizers
Functional compositionality: a G.N.U. approach
Physical and Mathematical Theory
Novel architectures and learning rules – oral contributions
Competitive Hebbian learning rule forms perfectly topology preserving maps
Approximating optimal information transmission using local Hebbian algorithms in a double feedback loop
Time-varying neural networks for large tasks
A "self-referential" weight matrix
Novel architecture and learning rules – poster contributions
Neural network complexity reduction using adaptive polynomial activation functions
FIELDNET, a dynamic network for pattern classification
Reducing the ratio between learning complexity and number of time varying variables in fully recurrent nets
Deletion of trained patterns by incremental learning in artificial neural network using Fahlman-Lebiere learning algorithm
Cascade neural network developed for time series prediction
On the information capacity of auto-associative RAM-based neural networks
Modified CMAC neural network architectures for nonlinear dynamic system modeling
Preliminary results on adaptively trained neural networks
Networks for learning and differentiating an input-output mapping
Design vs. training of neural machines
Counterexample of Witsenhausen under set-bounded model of uncertainty and its neural net solver
A modified learning algorithm for backpropagation network
High-order Boltzmann machines for MAX-SAT and SAT
EBP algorithm can work with hard limiters
Stochastic neural networks
A neurophysiologically motivated neural network model and its application to the superposition problem
A symmetrical lateral inhibition network for PCA and feature decorrelation
Minimizing the system error in feedforward neural networks with evolution strategy
Prove of convergence of extended divide and conquer networks
Monotonic incrementation of backpropagation networks
A multi-layer extension of a Bayesian neural network
YPROP: yet another accelerating technique for the back propagation
Automatic construction of multilayer networks for non linear regression
Generalization of a parametric learning rule
Supervised learning for decorrelated Gaussian networks
Associative memories that can form hypotheses: phase coded network architectures
A formal link between multilayer perceptrons and a generalization of linear discriminant analysis
Adaptive critic and probabilistic logic nets
DEFAnet2-advancements of a deterministic function approximator
Document retrieval and protein sequence matching using a neural network
A fuzzy neural architecture for supervised learning and classification of temporal sequences
Hierarchical reinforcement learning
Connectivity maximization of layered neural networks for supervised learning
The overlapped tessellaton: a supervised neural rule
!ABP: Interval Arithmetic Backpropagation
Architecture of associative memory with reduced cross talk and its performance formulation
Augmentation of generalisation in probabilistic logic nets
Fuzzy expert networks
Stochastic dynamical systems – oral contributions
Using Boltzmann Machines for probability estimation
Brownian motion updating of multi-layered perceptrons
Guaranteed convergence of learning in neural networks
Activity-conserving dynamics for neural networks
Stochastic dynamical systems – poster contributions
The lower bound of the capacity for a neural network with multiple hidden layers
A method for finding the optimal number of learning samples and hidden units for function approximation with a feedforward network
The N-2-N encoder: a matter of representation
Optimizing the architecture of multi-layer perceptrons for one-dimensional classification
Neural networks and genetic algorithms: improving the fault tolerance capabilities
Entropy of perceptrons
Assessing generalization by 2-D receptive field visualization
A fast training algorithm for feedforward neural networks
Improvement of the convergence of the learning using the modified back-propagation method
Selforganization – oral contributions
Parametrized self-organizing maps
Population dynamics on the basis of vector quantization: a method for auto-association and classification
Vector quantization with a growing and splitting elastic net
Learning topology-preserving maps using self-supervised backpropagation
Selforganization – poster contributions
A multiassociative memory for control
Phase transitions in self-organized feature maps
Unsupervised extraction of predictable abstract features
Genetic algorithm with migration on topology conserving maps
Analyzing Kohonen maps with geometry
A comparison between classical unsupervised classifiers and ART3 neural networks
A dynamic procedure for neural network design
PCA in a network with full lateral connections
Non-uniform cellular automata
SUSOM-"Supervised" Self-Organizing Maps
Dynamical systems – oral contributions
Synchrony in integrate-and-fire networks
A neural network for motion detection
Spikes or rates?-stationary, oscillatory, and spatio-temporal states in an associative network of spiking neurons
Cooperative stochastic effects in globally coupled bistable elements
Dynamical systems – poster contributions
Biologically inspired neural network for trajectory formation and obstacle avoidance
Catastrophic phase transitions in exact ART networks
Analysis of chaotic behaviour in dynamical systems using analog neural networks
A dynamically generalising weightless neural element
Vector quantization by neuro-dynamical system
The effect of synaptic time constants on firing patterns in populations of spiking neurons
Information processing by spatio-temporal chaotic networks
Hysteresis phenomena and bifurcation of periodic solutions in a mathematical model of cortical dynamics
Computing complexity of symmetric quadratic neural networks
Topology learning solved by extended objects: a neural network model
Higher order neural networks in a unified learning scheme
On a simple hysteresis network
Switching the vector field according to the input of an oscillatory neural network
Processing of information encoded in coupled one-dimensional maps
Feedback in single continuous neurons
A neural network for decision making in dynamic environments
Chaos in neural networks at nonlinear synapses
Stability conditions for nonlinear continuous random neural networks
Attractor neural networks – oral contributions
Optimal classification with multilayer networks
An attractor network model for the generation of event-related potentials using integrative synapses
Novel Liapunov functions for additive neural networks
Capacity and error correction ability of sparsely encoded associative memory with forgetting process
Attractor neural networks – poster contributions
Defining the attractor of a recurrent neural network by Boolean expressions
Using REDUCE for replica calculations
Equilibrium statistical mechanics of non-symmetric neural networks
Storage of words by coupling Hopfield nets
Constraints on learning in dynamic synapses
Recursive construction of neural networks with long periodic behavior
Phase-space gardening in the binary-couplings memory network
The relationship between choice of representation, network structure and performance in Harmony Theory networks
Learning and generalization – oral contributions
On the power of linearly weighted neural networks
Elimination of overtraining by a mutual information network
Cascade correlation: an incremental tool for function approximation
Learning and generalization – poster contributions
Bounds on the complexity of testing and loading neurons
Principal hidden unit analysis with minimum entropy method
Empirical criteria to compare the performance of neuro algorithms
LS-backpropagation algorithm for training multilayer perceptrons
Do backpropagation trained neural networks have normal weight distributions?
A constructive algorithm for binary mapping
BOXES revisited
Mathematical properties of multi-layer adaptive filters
Weight zero enhancement in speech synthesis using neural networks
Biological metaphors in designing modular artificial neural networks
Learning and generalization controlled by contradiction
Extraction of symbolic statements from synaptic weights
A novel back propagation algorithm with optimal number of hidden units
Two neural models for fast category learning-neural associative memories and the restricted Coulomb energy model
Storage capacity results for decomposed structures of generalizing RAM nodes
Applications
Industrial applications – oral contributions
Novelty detection and neural network validation
Estimating material properties for process optimization
Hybrid digital signal processing and neural networks for automated diagnostics using eddy current inspection
Industrial applications – poster contributions
Self-organizing neural network for diagnosis
Limitations of adaptive critic control schemes
Periodic disturbance rejection: a neural network approach
Representation of real-valued functions by a three-layered artificial neural network with topologically ordered input and output units
Prediction of reflectance values: towards the integration of neural and conventional colorimetry
Neural network modeling and prediction of multivariate time series using predictive MDL principle
Dynamics of a neural network-based financial market
Evolving neurocontrollers for pole balancing
Backpropagation vector quantization for satellite coverage plans optimization
Sequential self-organization for the traveling salesman problem
Invariant process control using neural networks
Optimal control of dynamic systems using self-organising maps
Monitoring a control system with a hybrid neural network architecture
Paper web profile and analysis using neural networks
Modelling of quality properties in paper drying with multilayer perceptron network
Interpolation of stationary non-linear time series by an optimized neural network
Two-sensor neural network modeling for fault detection
Modelling the fed batch fermentation process using artificial neural networks
Identification of car body steel by an optical on line system and a Kohonen's self-organizing map
Simulation of pulsed laser material processing controlled by an extended self-organizing Kohonen feature map
Process modelling using artificial neural networks
Real-time nuclear power plant monitoring with adaptively trained neural network
Self organized feature maps for monitoring and knowledge acquisition of a chemical process
Flow regime identification by a self-organising neural network
Functional electrical stimulation with neural network controlled state feedback
Artificial interacting agents for stock market experiments: the cross-target method
Neural network training by parameter optimization approach
Neural network analysis of the Hungarian party-state system
Modelling time-varying industrial processes using MLP networks
Pattern recognition I – oral contributions
Lithofacies identification from wireline logs-bringing neural networks to application
Using selforganizing feature maps to classify EEG coherence maps
Building an artificial retina for distance- and orientation-invariant pattern recognition
MLP-RBF: a cooperative multi-modular neural network application in high-energy physics
Pattern recognition I – poster contributions
Operational cloud classifier based on the topological feature map
Image segmentation using a self-organising logical neural network
High-resolution classification of Papanicolauo smear cells using back-propagation neural networks
Artificial neural networks detect subtle differences between anesthetics
Pattern segmentation and feature linking as simultaneous processes in an associative network of spiking neurons
Spatial topology distance for handprinted character recognition
Identification of underwater sonar images using fuzzy-neural architecture FuNe I
Fault detection in multivariate time series with a coding approach
Practical implementation of a radial basis function network for handwritten digit recognition
An efficient method of neural network application to recognizing of handwritten digits in ZIP codes
The application of average gradient matrices for fingerprint classification using neural networks
Neural architectures for motion tracking
A multi-agent classifier using associative networks in parallel
Handwritten alphabet and digit character recognition using skeleton pattern mapping with structural constraints
Optimization of a signature verification system using neural networks
Incremental case-based pattern classifier
Cepstral blur identification by neural network for image restoration purpose
Reduced pattern recognizing neural nets
Digit recognition by the random neural network using supervised learning Detecting abnormalities in MRl images using the difference method
Neural networks for the echographic diagnosis of diffuse liver diseases
Pattern recognition II – oral contributions
Neural networks and the travelling salesman problem
Automatically structured neural networks for handwritten character and word recognition
Tracking rain cells in radar images using multilayer neural networks
Neural network analysis of plasma spectra
Pattern recognition II – poster contributions
Monitoring EEG signal with self-organizing map
Invariant pattern recognition with recovery of transformation parameters
Applying dynamic link matching to object recognition in real world images
Performance of the backpropagation neural network for recognition of radio signals using time-domain features
Conceptual fuzzy sets application to facial expression recognition using associative memory system
Neocognitron with non-uniform receptive fields
Hand-written character recognition by a structured self-growing neural network "CombNET-II"
Segmentation of image sequences using self-organizing feature maps
Combining neural-network and statistical methods in seismic first-arrival picking
A study of neural network input data for ground cover identification in satellite images
On generalization ability of cascaded neural net architecture
Assessing the latency of peak Pa in auditory evoked potentials using neural networks
A self-organizing network of alterable competitive layer for pattern cluster
Prediction of secondary structures of proteins: comparison of neural networks (fuzzy ARTMAP) and statistical techniques
Cognitive grammar and map digitization
On-line learning with learning vector quantization: a case study of EEG classification
Image sequence coding using a neural vector quantization
Minimum distance pattern classifiers based on a new distance metric
Knowledge extraction by self organising maps
Application of the sensitivity algorithm in biological fields
Neural hardware and software – oral contributions
Challenge of ANN to microelectronics
Implementation of million connections neural hardward with URAN-I
Multiprocessor and memory architecture of the neurocomputer SYNAPSE-1
COKOS: A COprocessor for KOhonen's Selforganizing map
Neural hardware and software – poster contributions
Hardware implementation of Kohonen's feature map by scalar and SIMD-array processors
A nonlinear electronic layer for distributed neural nets
How to find a near optimal mapping of neural networks onto message passing multicomputers
20 million patterns per second VLSI neural network pattern classifier
High-density analog-EEPROM based neural network
A simple training law suitable for on-chip learning
Simulation of neural networks and genetic algorithms in a distributed computing environment using NeuroGraph
A parallel implementation of the back-propagation of errors learning algorithm on a SIMD parallel computer
CONVIS, a distributed environment for control and visualization of neural network simulation programs
Mapping of some neural network algorithms to a general purpose parallel neurocomputer
Architecture of low cost, large scale neural networks
A generalized recurrent neural network for matrix inversion
On the realization of back-propagation on a transputer based system
Self-organisation of large feature maps using local computations: analysis and VLSI integration
NEUROCOBOL: a COBOL-like neural network simulation language based on the layer macro definition
Encapsulated objects for neural network simulation
A harmony theory network solution to the N-Queens problem