Workshops

 

Machine learning in human language technology (W01) N. Fakotakis Schedule
Quality Assurance in Knowledge Discovery in Databases (W02) V. Moustakis  Tutorial
Machine learning in user modeling (W03) C. Papatheodorou Schedule
Machine learning for intelligent information access (W04) G. Karakoulas Schedule
Pre- and post-processing in machine learning and data mining: theoretical aspects and applications (W06) I. Bruha Schedule
Intelligent techniques for Spatio-Temporal Data Anaysis in Environmental Applications (W07) N. Avouris
N. Vassilas
Schedule
Data mining in economics, marketing and finance (W08) H. Jessen Schedule
Machine learning applications in the electric power industry (W09) N. Hatziargyriou Schedule
Support vector machine theory and applications (W10) T. Evgeniou Schedule
Machine learning and intelligent agents (W11) T. Panayiotopoulos Schedule
Biologically-inspired machine learning (W12) J. Demiris Schedule
Machine learning in medical applications (W13) G. Magoulas
A. Prentza
Schedule
Algorithmic Issues of the Genetic Approach (W14) P. Spirakis  Tutorial

Machine learning in human language technology - N. Fakotakis

July 7 - 8, 1999

DAY 1 - Wednesday, July 7, 1999

14:30 - 16:00 Invited talk "About Knowledge Discovery in Texts: A Definition and an Example"
Yves Kodratoff,
CNRS, France

16:00 - 16:30 Coffee Break

16:30 - 16:50 G.Orphanos
Decision Trees and NLP: A Case Study in POS Tagging
G.Orphanos, D.Kalles, T.Papagelis, D.Christodoulakis
Patras University & Computer Technology Institute, Patras, Greece

16:50 - 17:10 B.Megyesi
Brill's PoS Tagger with Extended Lexical Templates for Hungarian
B.Megyesi
Stockholm University, Sweden

17:10 - 17:30 G. Petasis
Part-Of-Speech Ambiguity in the Greek Language Using Learning Techniques
G.Petasis, G.Paliouras, V.Karkaletsis, C.D.Spyropoulos, I.Androutsopoulos
N.C.S.R. "Demokritos", Athens, Greece

17:30 - 18:00 Discussion

DAY 2 - Thursday, July 8, 1999

14:30 - 14:50 D. Levinson
Corpus-Based Method for Unsupervised Word Sense Disambiguation
D.Levinson
Israel Institute of Technology

14:50 - 15:10
Lexical Learning for Improving Syntactic Analysis
R.Basili, M.T.Pazienza, M.Vindigni
Univ. of Roma

15:10 - 15:30
Unknown Word Acquisition System for Japanese Written-Language Document
H.Kameda, T.Sakurai, C.Kubomura
Tokyo Univ. of Technology, Japan

15:30 - 15:50 B.Thomas
Learning T-Wrappers for Information Extraction
B.Thomas
Univ. Koblenz, Germany

15:50 - 16:00 Discussion

16:00 - 16:30 Coffee Break

16:30 - 16:50
Automatic Modeling of a Task for Realizing a Multi-modal Interface System
S.Sakurai, T.Endo, T.Mukai, R.Oka
Real World Computing Partnership, Japan

16:50 - 17:10 E. Stamatatos
Automatic Extraction of Rules for Sentence Boundary Disambiguation
E.Stamatatos, N.Fakotakis, G.Kokkinakis
University of Patras, Greece

17:10 - 17:30 J.Veenstra
Memory-Based Text Chunking
J.Veenstra
Tilburg University, the Netherlands


Quality Assurance in Knowledge Discovery in Databases - V. Moustakis

Tuesday, July 6, 1999

Tutorial by:
Prof. Vassilis Moustakis
Technical University of Crete, Chania, Greece

This tutorial couples principles, methods and techniques drawn from quality management with Knowledge Discovery in Databases (KDD) specifics. A major issue in KDD applications is the quality of the derived output, which may be expressed in tree or rule format, or in other representations. The ultimate success of KDD lies in the degree of user confidence with respect to the derived output. However, such confidence is difficult to be achieved given that KDD are based on inductive inference known to be falsity preserving rather than truth preserving.

The tutorial will adopt a case format. Alternative KDD or ML applications will be presented and discussed aiming towards motivating participants on the issues involved. Furthermore, the tutorial aims to bridge the gap between two technical/scientific fields: ML and quality management.


Machine learning in user modeling - C. Papatheodorou

Tuesday, July 6, 1999

14:30 - 15: 20 Invited Talk: User Modeling in Adaptive Interfaces
P. Langley
AT& T Labs, USA

15:20 - 15:40 Learning User Profiles from Positive Examples
Ingo Schwab and Wolfgang Pohl
German National Research Center for Information Technology (GMD), Germany

15:40 - 16:00 Machine Learning Techniques for Adaptive User Interfaces in a Corporate Digital Library Service
G. Semeraro, M.F. Costabile, F. Esposito, N. Fanizzi, S. Ferilli
Univ. of Bari, Italy

16:00 - 16:30 Coffee Break

16:30 - 16:50 User Modeling in Information Discovery: An Overview
D. Fragoudis, S. Likothanassis
University of Patras, Greece

16:50 - 17:10 User Modeling in a MultiAgent Evolving System
Alex Moukas
MIT, USA

17:10 - 17:30 Dual-model: An Architecture for Utilizing Temporal Information in Student Modeling.
Bark Cheung Chiu and Geoffrey I. Webb
School of Computing and Mathematics, Deakin University, Victoria, Australia

17:30 - 17:50 A Data Miner analyzing Navigational Behavior of Web Users
M. Spiliopoulou, L. Faulstich, K. Winkler
Humboldt University, Berlin, Germany

17:50 - 18:10 Learning Communities of the ACAI '99 Web Site Visitors
G.Paliouras, C. Papatheodorou, V. Karkaletsis, P. Tzitziras, C.D. Spyropoulos
NCSR "Demokritos", Athens, Greece

18:10 - 18:30 Discussion


Machine learning for intelligent information access - G. Karakoulas

Wednesday, July 7, 1999

14:30-14:40 Introducing the workshop
G. Semeraro
Univ. of Bari, Italy

14:40-15:10
Comparing and combining retrieval feedback methods: first results
C. Carpineto and G. Romano
Fondazione Ugo Bordoni, Italy

15:10-15:35
Dynamic aspects in information filtering: detecting changes in text streams
C. Lanquillon
Daimler-Chrysler, Ulm, Germany

15:35-16:00
Supporting document acquisition and organization in a digital library service through ML techniques
F. Esposito, N. Fanizzi, S. Ferilli and G. Semeraro
Univ. of Bari, Italy

16:00-16:30 Coffee Break

16:30-16:55
Predicting relevance of documents with a memory-based information agent
D. D'Aloisi and V. Giannini
Fondazione Ugo Bordoni, Italy

16:55-17:20
Intelligent information access in the web: ML based user modeling for high precision meta-search
M. Muller
University of Osnabrueck, Osnabrueck, Germany

17:20-18:00 Discussion


Pre- and post-processing in machine learning and data mining: theoretical aspects and applications - I. Bruha

Thursday, July 8, 1999

14:30 - 15:15 Survey talk
From Machine Learning to Knowledge Discovery: preprocessing and postprocessing
I.Bruha
McMaster Univ., Canada

15:15 - 15:35 T. Brijs
Reducing redundancy in characteristic rule discovery by using IP-techniques.
Tom Brijs, Koen Vanhoof, Geert Wets
Limburg Univ., Fac. Applied Economic Sciences, Belgium

15:35 - 15:55 D. Kalles
Induction of decision trees in numeric domains using set-valued attributes.
D. Kalles, A. Papagelis
Computer Technology Institute, Patras, Greece

16:00 - 16:30 break

16:30 - 16:50 J. Lahtinen
Using Bayesian networks for visualizing high-dimensional data.
P. Kontkanen, J.Lahtinen, P.Myllymaki, T.Silander, H.Tirri
Complex System Computation Group, Univ of Helsinki, Finland

16:50 - 17:10 J. Smid
Web-based data analysis tool.
J. Smid, P.Svacek, P.Volf, E.Levine, L.Kurz, I.Gluksmann, M.Smid
Morgan State Univ, USA

17:10 - 17:30 L. Talavera
Dependancy-based dimensionality reduction for clustering symbolic data.
L. Talavera,
Dept de Llenguatges i Sistemes Informatics, Univ Plitechnica, Catalunya, Spain

17:30 - 17:50 I. Bruha
Discretizing numerical attributes in a genetic attribute-based learning algorithm.
P. Kralik, Technical Univ Brno, Czech Rep.
I. Bruha, McMaster Univ., Canada

17:50 - 18:10 Discussion, Conclusion


Intelligent techniques for Spatio-Temporal Data Anaysis in Environmental Applications - N. Avouris, N. Vassilas

Tuesday, July 13, 1999

14:30 - 14:45 Welcome Speech, Introduction
[Organizing Committee]

14:45 - 15:00 Intelligent Fusion of Multisource Data for Sea Ice Classification
Costas Tsatsoulis, Leen-Kiat Soh, Cheryl Bertoia and Kim Partington
University of Kansas, USA

15:00 - 15:15 Fusion and Classification of Multitype Data
Jon Atli Benediktsson and Kolbeinn Arnason
University of Iceland, Iceland

15:15 - 15:30 Automatic Lineament Detection from Geophysical Grid Data Using Efficient Clustering and Weighted Hough Transform Algorithms
N. Vassilas, S. Perantonis, E. Charou, K. Seretis and Th. Tsenoglou
NCSR "Demokritos", Athens, Greece

15:30 - 15:45 ISLE: ISLE: An Intelligent System for Land Evaluation
G.Tsoumakas and I.Vlahavas
Aristotle University of Thessaloniki, Greece

15:45 - 16:00 Decision-Oriented Environmental Mapping with Radial Basis Function Neural Networks
V. Demyanov, N. Gilardi, M. Kanevski, M. Maignan and V. Polishchuk
Institute of Nuclear Safety (IBRAE), Russia

16:00 - 16:30 Coffee Break

16:30 - 16:45 Environmental and Pollution Spatial Data Classification with Support Vector Machines and Geostatistics
N. Gilardi, M. Kanevski, M. Maignan and E. Mayoraz
IDIAP Research Institute, Switzerland

16:45 - 17:00 Machine Learning Techniques for Air Quality Prediction
Elias Kalapanidas, Nikos Avouris
University of Patras, Greece

17:00 - 17:15 Tree-Based Classification Approach for Dealing with Complex Knowledge in Natural Sciences
David Grosser and Noel Conruyt
University of La Reunion, France

17:15 - 17:30 Bayesian Networks for Change Analysis in Environmental Monitoring
Jose Carlos Ferreira da Rocha, Dibio Leandro Borges and Celso Antonio Alves Kaestner
CEFET-PR and UEPG, Brasil

17:30 - 17:45 Bayesian Modeling in Paleoenvironmental Reconstruction
Hannu Toivonen, Kari Vasko, Heikki Mannila, Atte Korhola and Heikki Olander
University of Helsinki, Finland

17:45 - 18:00 Discussion / Conclusions


Data mining in economics, marketing and finance - H. Jessen

Tuesday, July 13, 1999

Knowledge Discovery in Practice
Ad Feelders, Hennie Daniels
Tilburg University, the Netherlands
Erasmus University Rotterdam, Rotterdam, the Netherlands

FIGI: Using Mobile Agent Technology to Collect Financial Information on Internet
Marios Dikaiakos
University of Cyprus, Cyprus

Stock Miner: A System for Knowledge Discovery in Financial Data
Melina Karamanlidou, Olivier Tuffier, Ioannis Vlahavas
Aristotle University of Thessaloniki, Greece
University of Nantes, France

16:00 - 16:30 Coffee Break

Analyzing business databases with the ProbRough rule induction system
Zdzislaw Piasta
Kielce University of Technology, Poland

Evolutionary Data Mining applied to TV Databases: A First Approach
Panagiotis Adamidis, Konstantinos Koukoulakis
Technological Education Inst. of Thessaloniki, Greece

The Improvement of Response Modelling: Combining Rule-Induction and Case-Based Reasoning
Filip Coenen, Gilbert Swinnen, Koen Vanhoof, Geert Wets
Limburg University Centre, Diepenbeek, Belgium

A fuzzy rule based learning method for corporate bankruptcy prediction
Nikolaos Thomaidis, George Dounias, Costas D. Zopounidis
Technical University of Crete, Chania, Greece


Machine learning applications in the electric power industry - N. Hatziargyriou

13 - 14 July 1999

DAY 1 - Tuesday, 13 July 1999

14:30 - 14:40
Introduction to the Workshop
N.D. Hatziargyriou, Department of Electrical and Computer Engineering, National Technical University of Athens, Athens

14:40 - 15:05
Supplementary Neural Controls for Improved Transient Behavior of Power Systems
D.J. Sobajic, Power Delivery Department, EPRI, Palo Alto, CA, R. Shen, B. Widrow, Stanford University, Stanford, CA, USA

15:05 - 15:30
Pruning Kernel Regression Trees for Security Assessment of the Crete Network
M. Helena Vasconcelos and J.A. Pecas Lopes, INESC, FEUP, Porto, Portugal

15:30 - 16:00
Dynamic Security Assessment by Fuzzy Inference
M. Matos, J.A. Pecas Lopes, M. Helena Vasconcelos, INESC, FEUP ,Porto, Portugal

16:00 - 16:30 Coffee Break

16:30 - 17:00
Genetic Algorithm Solution to the Economic Dispatch Problem - Application to the Electrical Power Grid of Crete Island
A. Saramourtsis, J. Damousis, A. Bakirtzis, P. Dokopoulos, Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, Thessaloniki

17:00 - 17:30
Probabilistic Neural Networks for Power Flow Analysis
A.E. Gavoyiannis, N. Vlassis, N.D. Hatziargyriou, Department of Electrical and Computer Engineering, National Technical University of Athens, Athens

17:30 - 18:00 Discussion

DAY 2 - Wednesday, 14 July 1999

14:30 - 15:00 D. Kalles
Intelligent Monitoring and Maintenance of Power Plants
D. Kalles, A. Stathaki and R. E. King, Computer Technology Institute, Department of Electrical & Computer Engineering, University of Patras, Patras

15:00 - 15:30
Machine Learning Applications in the Transformer Manufacturing Industry
P. Georgilakis, N. Hatziargyriou, Schneider electric AE, Elvim Plant, Department of Electrical and Computer Engineering, National Technical University of Athens, Athens

15:30 - 16:00
A Genetic Fuzzy Model for Wind Speed Prediction and Power Generation in Wind Parks
I.G. Damousis, P.D. Dokopoulos, Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, Thessaloniki

16:00 - 16:30 Coffee Break

16:30 - 17:00
Mesh Density Prediction for Open Boundary Faulted Power Transmission Line Problems Using Artificial Neural Networks
Dimitris Triantafyllidis, Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, Thessaloniki

17:00 - 17:30
An Innovative Genetic Fuzzy System for the Prediction of an Inductive Interference Problem
K.J. Satsios, I.G. Damousis, D.P. Labridis, P.S. Dokopoulos, Electrical and Computer Engineering Department, Aristotle University of Thessaloniki, Thessaloniki

17:30 - 18:00
Artificial Intelligence Approaches to the Problem of Power System Restoration
N. Fountas, N. Hatziargyriou, Department of Electrical and Computer Engineering, National Technical University of Athens, Athens

18:00 - 18:30 Discussion


Support vector machine theory and applications - T. Evgeniou

Wednesday, July 14, 1999

14:30-15:00 Nicolas Vayatis
How to estimate the Vapnik-Chervonenkis Dimension of Support Vector Machines through simulations
Nicolas Vayatis and Robert Azencott
Centre de Mathematiques et de Leurs Applications (CMLA)
Ecole Normale Superieure de Cachan, Cedex, France
and
Centre de Recherche en Epistemologie Appliquee (CREA)
Ecole Polytechnique, Cedex, France

15:00-15:30 Anastasios Tefas
Enhancing the performance of elastic graph matching for face authentications by using support vector machines
Anastasios Tefas, Constantine Kotropoulos and Ioannis Pitas
Department of Informatics, Aristotle University of Thessaloniki, Greece

15:30 - 16:00 K. Veropoulos
The Application of Support Vector Machines to Medical Decision Support: A Case Study
K. Veropoulos, N. Cristianini and C. Campbell
Department of Engineering Mathematics, Bristol University, Bristol, UK

16:00-16:30 Coffee break

17:00-17:30 Rodrigo Fernandez
Predicting time series with a local support vector regression machine
Rodrigo Fernandez
LIPN, Institut Galilee-Universite Paris 13, France

17:30-18:00 Theodore Trafalis
Primal-dual optimization methods in neural networks and support vector machines training
Theodore Trafalis
School of industrial engineering, University of Oklahoma, USA


Machine learning and intelligent agents - T. Panayiotopoulos

Thursday, July 15, 1999

14:30 - 14:45 Welcome Speech, Introduction
[Organizing Committee]

An information-theoretic localization criterion for robot map building
Nikos Vlassis        Yoichi Motomura             Ben Krose
RWCP, Autonomous Learning Functions SNN    Electrotechnical Laboratory
University of Amsterdam, The Netherlands Umezono 1-1-4, Japan

Learning to Envision: An Intelligent Agent for Ship Damage Control
Vadim V. Bulitko, David C. Wilkins
Beckman Institute, Univ. of Illinois at Urbana-Champaign, USA

Agents Learning in a Three-Valued Logical Setting
Evelina Lamma Fabritzio Riguzzi    Luis Moniz Pereira
DEIS, Universita di Bologna    Centro de Inteligencia Artificial (CENTRIA)
40136 Bologna, Italy, Departmento de Informatica, Uiversidade Nova de Lisboa, Portugal

16:00-16:30 Coffee break

Machine Learning used by Personal WebWatcher
Dunja Mladenic
Department of Intelligent Systems, J.Stefan Institute, Ljubljana, Slovenia

A Learning Natural Language Interface to an Agent
John Kontos and Ioanna Malagardi
Artificial Intelligence Laboratory, Department of Informatics, Athens University of Economics and Business

A Learning Personalized Information Agent for the WWW
Zacharis Nick, Themis Panayiotopoulos
University of Piraeus, Department of Computer Science, Piraeus, Greece


Biologically-inspired machine learning - J. Demiris

Wednesday, July 14, 1999

14:30 - 14:35 Introduction to the workshop
John Demiris and Gert Westermann
The University of Edinburgh, Scotland, UK

14:35 - 15:00 On the Implications of Delay Adaptability for Learning in Pulsed Neural Networks
Michael Schmitt
LS Mathematik & Informatik, Fakultaet fuer Mathematik, Ruhr-Universitaet Bochum, Germany

15:00 - 15:25 Image Categorization using Independent Component Analysis
Abderrahim Labbi, Holger Bosch and Christian Pellegrini
Department of Computer Science, University of Geneva, Switzerland

15:25 - 15:50 Learning to Navigate: from animal to robot
G. Li, B. Svensson

15:50 - 16:00 Using bat bio-sonar as a biological approach for mobile robots navigation
Jose Carmena and John Hallam,
Institute for Perception, Action and Behaviour,
Division of Informatics, University of Edinburgh, Scotland, UK

16:00 - 16:30: Coffee break

16:30 - 16:40 Evolution­based Optimization Models of Complex Interacting Systems
D. Solow, A. Burnetas, M. Tsai, N. Greenspan
Case Western Reserve University, Cleveland, USA

16:40 - 17:05 Preliminary approaches to attention for social learning
Yuval Marom and Gillian Hayes
Institute for Perception, Action and Behaviour,Division of Informatics, The University of Edinburgh, Scotland, UK

17:05 - 17:30 Selective Attention for Noise Robust Speech Recognition
Ki-Young Park and Soo-Young Lee
Brain Science Research Center -/- Department of Electrical Engineering
Korea Advanced Institute of Science and Technology, Korea (South)

17:30 - 18:00: Panel Discussion


Machine learning in medical applications - G. Magoulas, A. Prentza

Thursday, July 15, 1999

14:30 - 14:40 Introduction: Systems and Sensors for improved minimally invasive surgery
Dr. Marc O. Shurr,
Eberhard-Karls-University of Tuebingen, Germany

14:40 - 15:00 Machine Learning and Medical Decision Making
V. Moustakis, G. Charissis
Technical University of Crete, Chania, Greece

15:00 - 15:15 Medical Diagnosis of Stroke Using Inductive Machine Learning
E.Alexopoulos, G. D. Dounias, K. Vemmos
Technical University of Crete, Chania, Greece

15:15 - 15:30 Impact of Machine Learning to the Diagnosis and Prognosis of First Cerebral Paroxysm
I. Zelic, N. Lavrac, P. Najdenov, Z. Rener-Primec
J. Stefan Institute, Ljubljana, Slovenia
University Medical Center, Ljubljana, Slovenia

15:30 - 15:45 Computer Interactive System for Ascertainment of Visual Perception Disorders
R. Ruseckaite
Vytautas Magnus University, Kaunas, Lithuania

15:45 - 16:00 A knowledge acquisition and management system for ECG diagnosis
Ph. Bourlas, E. Giakoumakis, G. Papakonstantinou
Athens University of Economics & Business, Greece
Technical Univ. of Athens, Greece

16:00 - 16:30 Coffee Break

16:30 - 16:45 An Information System That Supports Knowledge Discovery and Data Mining in Medical Imaging
J. Neves, V. Alves, L. Nelas, A. Romeu, S. Basto

16:45 - 17:00 Identification of Individualized Models of the Human Cardiovascular System
A. Asteroth
Bonn University, Germany

17:00 - 17:15 Finding Similarities Between an Activity of the Different EEG's by means of a Single Layer Perceptron
Erinija Pranckeviciene

17:15 - 17:30 Approximation and Classification in Medicine with IncNet Neural Networks
N. Jankowski
Nicholas Copernicus University in Torun, Poland

17:30 - 17:45 Detecting abnormalities in endoscopic images by textural description and neural networks
S. Karkanis, G.D. Magoulas, M. Grigoriadou, M. Schurr
University of Athens, Greece

17:45 - 18:00 Classification of endoscopic images based on Texture Spectrum
S. Karkanis, K. Galoussi, D. Maroulis
University of Athens, Greece

18:00 - 18:30 Discussion / Conclusions


Algorithmic Issues of the Genetic Approach - P. Spirakis

Thursday, July 15, 1999

Tutorial by:
Prof. P. Spirakis and Dr. P.Psycharis,
University of Patras, Greece

The workshop aims to show that the genetic heuristic can in some cases be trusted to have a sound and complete mathematical support that lifts it from the set of heuristics to the case of probabilistic algorithms with guaranteed behavior. Applications to lottery problems and to stock market problems will be considered. Areas of interest include:

  1. Mathematical Foundations of Genetic Algorithms
  2. Quadratic dynamic systems and forecast of their behavior
  3. The problem of uniform random generation of combinatorial structures
  4. A genetic algorithm for random uniform generation of matchings of a graph
  5. The efficient parallel implementation of the genetic system
  6. Applications to stock market rules
  7. A comparison with linear systems (Markov chains)
  8. Open issues