WSOM'97 Scientific Program

---

WEDNESDAY, June 4

---

Tutorial Short Course (14:00-17:00, room G)

Self-organizing maps: state-of-the-art
Teuvo Kohonen - Helsinki University of Technology, Finland

---

THURSDAY, June 5

---

Opening (8:30-8:50, room G)

Teuvo Kohonen

---

Plenary session I (8:50-9:50, room G)
Chairperson: Teuvo Kohonen, Finland

Learning with the parameterized self-organizing map
Helge Ritter - University of Bielefeld, Germany

---

Session F1: Pattern recognition and signal processing (10:15-12:15, room F)
Chairpersons: Helge Ritter, Germany; Dieter Merkl, Austria

Self-organizing maps of symbol strings with application to speech recognition
Teuvo Kohonen and Panu Somervuo - Helsinki University of Technology, Finland

SOM based density function approximation for mixture density HMMs
Mikko Kurimo - Helsinki University of Technology, Finland

Modelling and data analysis of multisensor systems with the self-organizing map: application to the electronic nose
Corrado Di Natale and Arnaldo D'Amico - University of Rome, Italy

Neural detection of QAM signal with strongly nonlinear receiver
Kimmo Raivio - Helsinki University of Technology, Finland, Jukka Henriksson - Nokia Research Center, Finland, Olli Simula - Helsinki University of Technology, Finland

Self-organization and segmentation with laterally connected maps of spiking neurons
Yoonsuck Choe and Risto Miikkulainen - The University of Texas at Austin, USA

Local subspace classifier and local subspace SOM
Jorma Laaksonen - Helsinki University of Technology, Finland

---

Session G1: Monitoring and data mining (10:15-12:15, room G)
Chairpersons: Karl Goser, Germany; Heizo Tokutaka, Japan

The use of neural networks in marketing: market segmentation with self organising feature maps
Sara Dolnicar - Vienna University of Economics and Business Administration, Austria

New approaches of process visualization and analysis in power plants
Ralf Otte - ABB Kraftwerksleittechnik, Germany, Karl Goser - University of Dortmund, Germany

Using the SOM in supporting diabetes therapy
Mikko Mäkipää - Helsinki University of Technology, Finland, Pekka Heinonen - Nokia Mobile Phones, Finland, Erkki Oja - Helsinki University of Technology, Finland

Scaling and sensitivity in appraisal
Eero Carlson - National Land Survey of Finland

CorDex: a knowledge discovery tool
Robert Leivian, William Peterson and Mike Gardner - Motorola, USA

The neural data analysis environment
Erkki Häkkinen and Pasi Koikkalainen - University of Jyväskylä, Finland

---

Plenary session II (13:30-14:30, room G)
Chairperson: Olli Simula, Finland

Self-organizing map for intelligent process control
Karl Goser - University of Dortmund, Germany

---

Session F2: Optimization and Poster spotlights (14:30-15:10, room F)
Chairpersons: Visa Koivunen, Finland; Jukka Heikkonen, Finland

The optimization for TSP using SOM method of many cities, for example 532 cities in USA
Kikuo Fujimura, Heizo Tokutaka, Shin-ichi Tanaka, Takatoshi Maeno, and Satoru Kishida - Tottori University, Japan

Poster spotlights (2 minutes each):

A hierarchical self-organizing map for classification problems
Bart De Ketelaere, Dimitrios Moshou and Peter Coucke - K.U. Leuven, Belgium

An angular quantising self organising map for scale invariant classification
Stephen McGlinchey and Colin Fyfe - University of Paisley, UK

Self-organizing maps for independent component analysis
Petteri Pajunen and Juha Karhunen - Helsinki University of Technology, Finland

Projection pursuits in SOM classifiers
V. Chandrasekaran and Zhi-Qiang Liu - University of Melbourne, Australia

Alternative ways for cluster visualization in self-organizing maps
Dieter Merkl and Andreas Rauber - Technical University of Vienna, Austria

On the merits of topography in neural maps
Michael Herrmann - Max-Planck-Institut für Strömungsforschung, Germany

Comparison of a Bayesian SOM with the EM algorithm for Gaussian mixtures
Hujun Yin and Nigel M. Allinson - UMIST, UK

Minimum description length (MDL) principle and self-organizing maps
Heikki Hyötyniemi - Helsinki University of Technology, Finland

A modular approach to implementation of the self-organising map
N. Lightowler, C.T. Spracklen, and A.R. Allen - University of Aberdeen, UK

---

Session G2: Hardware and Poster spotlights (14:30-15:10, room G)
Chairpersons: Pasi Koikkalainen, Finland; Krista Lagus, Finland

SOM hardware-accelerator
S. Rüping, M. Porrman and U. Rückert - University of Paderborn, Germany

Poster spotlights (2 minutes each):

Data parallel simulation of self-organizing maps on hypercube architectures
Erich Schikuta and Claus Weidmann - University of Vienna, Austria

Object localisation in 2d images using a temporal Kohonen network
H.M. Lakany and G.M. Hayes - University of Edinburgh, UK

MuSeq recurrent oscillatory self-organizing map. Classification and entrainment of temporal feature spaces
Mauri Kaipainen - Indiana University, USA, Pantelis Papadopoulos - Indiana University, USA, Pasi Karhu - University of Helsinki, Finland

Classification of human blood plasma lipid abnormalities by 1H magnetic resonance spectroscopy and self-organizing maps
Yrjö Hiltunen - The Raahe Institute of Computer Engineering, Finland, Jouni Kaartinen - The Raahe Institute of Computer Engineering, Finland, Mika Ala-Korpela - University of Kuopio, Finland

Coordinate-free representation of sensorimotor spaces
Francesco Frisone, Pietro G. Morasso, Vittorio Sanguineti - University of Genova, Italy

WarpNet: self-organizing time warping
Kari Torkkola - Motorola, USA

Dynamic neighbourhoods in self organizing maps
Jean-Jacques Mariage - Paris 8 University, France

Extended self-organizing maps with local linear mappings for function approximation and system identification
Dimitrios Moshou and Herman Ramon - K.U. Leuven, Belgium

Self organizing evolutionary models in financial markets forecasting
Marina Resta - Genova University, Italy

---

Poster and demo session (15:10-16:20, Lobby)

---

Session F3: Temporal sequence processing (16:20-17:40, room F)
Chairpersons: Thomas Martinez, Germany; Jari Kangas, Finland

The adaptive-subspace self-organizing map (ASSOM)
Teuvo Kohonen, Samuel Kaski, Harri Lappalainen, and Jarkko Salojärvi - Helsinki University of Technology, Finland

Context learning with the self organizing map
Markus Varsta and Jukka Heikkonen - Helsinki University of Technology, Finland, Jose del R. Millan - Joint Research Centre of the European Commission, Italy

Local linear correlation analysis with the SOM
Antonio Piras and Alain Germond - Swiss Federal Institute of Technology, Switzerland

Using the SOM and local models in time-series prediction
Juha Vesanto - Helsinki University of Technology, Finland

---

Session G3: Image analysis and vision (16:20-17:40, room G)
Chairpersons: Risto Miikkulainen, USA; Jouko Lampinen, Finland

Application of a growing self-organizing map to thinning of binary characters with noise
Igor Farkas and Lucius Chudy - Slovak Academy of Sciences, Slovakia

Towards self-organized feature maps from Gabor filter responses
James Orwell - Kings College London, UK, Ramón Turnes, María José Carreira and Diego Cabello - Univ. Santiago, Spain, James Boyce - Kings College London, UK

TEXSOM: A new architecture for texture segmentation
Javier Ruiz-del-Solar - Fraunhofer-Institut IPK Berlin, Germany

Analysis of SOM-based models for the development of visual maps
H.-U. Bauer - MPI Strömungsforschung, Germany, M. Riesenhuber - MIT, USA, D. Brockmann and T. Geisel - MPI Strömungsforschung, Germany

---

Banquet and dinner speech (18:00 - 22:30, Fortress Island of Suomenlinna)

A neural network saga
Robert Hecht-Nielsen - HNC Software Inc. and UCSD, USA

---

FRIDAY, June 6

---

Plenary session III (8:30-9:30, room G)
Chairperson: Erkki Oja, Finland

Theoretical aspects of the SOM algorithm
Marie Cottrell - Univ. Paris I, France

---

Session F4: Invited session (9:50-10:20, room F)
Chairpersons: Marie Cottrell, France; Heikki Hyötyniemi, Finland

SOM research in the USA
Risto Miikkulainen - The University of Texas at Austin, USA

---

Session F5: Theory and extensions (10:20-12:00, room F)
Chairpersons: Marie Cottrell, France; Heikki Hyötyniemi, Finland

Self-organisation in the one-dimensional SOM with a reduced width neighbourhood
John A. Flanagan - Swiss Federal Institute for Technology, Switzerland

A comparison of topography measures for neural maps
M. Herrmann and H.-U. Bauer - Max-Planck-Institut für Strömungsforschung, Germany, Th. Villmann - Universität Leipzig, Germany

Organization measures for self-organizing maps
Daniel Polani - Johannes Gutenberg-Universität Mainz, Germany

The GSOM-algorithm for growing hypercubical output spaces in self-organizing maps
Thomas Villmann - University Leipzig, Germany, H.-U. Bauer - Max-Planck-Institut für Strömungsforschung, Germany

Two more modifications of SOMs to handle signals with special properties
Stefan Schuenemann, Udo Seiffert, Bernd Michaelis - Otto-von-Guericke-Universität Magdeburg, Germany

---

Session G4: Text and document maps (10:00-11:20, room G)
Chairpersons: Xia Lin, USA; Barbro Back, Finland

Comparisons of self-organized word category maps
Timo Honkela - Helsinki University of Technology, Finland

Experiments on distributional categorization of lexical items with self organizing maps
Gilles Bernard - Paris 8 University, France

WEBSOM - self-organizing maps of document collections
Timo Honkela, Samuel Kaski, Krista Lagus, and Teuvo Kohonen - Helsinki University of Technology, Finland

Lessons learned in text document classification
Dieter Merkl - Technische Universität Wien, Austria

---

Session G5: Invited session (11:20-11:50, room G)
Chairpersons: Xia Lin, USA; Barbro Back, Finland

Condensed review of SOM and LVQ research in Japan
Heizo Tokutaka - Tottori University, Japan

---

Panel: Document search (13:00-14:10, room G)

Participants:

Robert Hecht-Nielsen, USA

Timo Honkela, Finland

Xia Lin, USA

Dieter Merkl, Austria

Jakub Zavrel, Netherlands

---

Session F6: Probabilistic interpretations (14:10-15:10, room F)
Chairpersons: Nigel Allinson, UK; Ari Hämäläinen, Finland

Magnification factors for the SOM and GTM algorithms
Christopher M. Bishop, Markus Svensen, and Christopher K. I. Williams - Aston University, UK

Self organizing map, a probabilistic approach
F. Anouar and F. Badran - Conservatoire National des Arts et Métiers (CEDRIC), France, S. Thiria - Université de Paris, France

Deterministic annealing for topographic vector quantization and self-organizing maps
Thore Graepel, Matthias Burger, and Klaus Obermayer - Technical University of Berlin, Germany

---

Session G6: Financial analysis (14:10-15:10, room G)
Chairpersons: Guido Deboeck, USA; Samuel Kaski, Finland

Neural network analysis of Russian banks
S. A. Shumsky and A. V. Yarovoy - Lebedev Physics Institute, Russia

Analyzing financial performance with self-organizing maps
Barbro Back - Turku School of Economics and Business Administration, Finland, Kaisa Sere - Åbo Akademi University, Finland, Hannu Vanharanta - Lappeenranta University of Technology, Finland

Analyzing financial statements with the self-organizing map
Kimmo Kiviluoto - Helsinki University of Technology, Finland, Pentti Bergius - Kera Ltd., Finland

---

Last modified: Mon Jun 2 17:36:07 EET DST 1997