Soft Computing in Case Based Reasoning

Springer London Berlin Heidelberg New York Barcelona Hong Kong Milan Paris Singapore Tokyo

Sankar K. Pal, Tharam S. Dillon and Daniel S. Yeung (Eds)

Soft Computing in Case Based Reasoning

i Springer

Sankar K. Pal, MTech, PhD, DIC, Fellow IEEE Distinguished Scientist and Head, Machine Intelligence Unit, Indian Statistical Institute, 203 Barrackpore Trunk Road, Calcutta 700 035, India

Tharam S. Dillon, BE, PhD. FSaRS, FIE, SMIEEE Department of Computer Science and Computer Engineering, La Trobe University, Bundoora, Melbourne, Victoria 3083, Australia

Daniel S. Yeung, BA, MA, MSc, MBA, PhD, FHKIE Department of Computing, The Hong Kong Polytechnic University, Hunghom, Kowloon, Hong Kong

ISBN-13: 978-1-85233-262-4

DOl: 10.1007/978-1-4471-0687-6

e-ISBN-13: 978-1-4471-0687-6

British Library Cataloguing in Publication Data Soft computing in case based reasoning

1. Soft computing. 2. Case-based reasoning I. Pal, Sankar K. II. Dillon, Tharam S. m. Yeung, Daniel S. 006.3'33

ISBN 185233262X

Library of Congress Cataloging-in-Publication Data Soft computing in case based reasoning I Sankar K. Pal, Tharam S. Dillon and Daniel S. Yeung (eds.),

p.em. Includes bibliographical references.

1. Soft computing. 2. Case-based reasoning. I. Pal, Sankar K. II. Dillon, Tharam S., 1943- m. Yeung, Daniel S., 1946-QA76.9.563 5625 2000 006.3-dc21

00-032973

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced. stored or transmitted. in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms of licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be sent to the publishers.

© Springer-Verlag London Limited 2001

Softcover reprint of the hardcover 1st edition 2001

The use of registered names, trademarks etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore free for general use.

The publisher makes no representation, express or implied. with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility or liability for any errors or omissions that may be made.

Typesetting: Camera ready by editors

-3413830-543210 Printed on acid-free paper SPIN 10755259

To the co-authors of all my research publications

- Sankar

To my parents Gurdial Singh and Kartar Kaur Dillon

for their constant support, tolerance and unselfish love

- Tharam

To Jesus Christ, my Lord and God and Foo Lau, my wife and best friend

- Daniel

Foreword

When, at the beginning of the 90s, the three areas of fuzzy set theory, artificial neural nets and genetic algorithms joined forces under the roof of soft computing or com­putational intelligence, people considered as their common feature the origin of the three areas; namely, the observation of biological systems and the mimesis of certain features which were promising to improve human reasoning, problem solving, diag­nosis or decision making. The three areas were considered to be complementary and synergies were expected. These hopes have turned out to be justified. It was, however, not sufficiently realised that another of the common features is that all three paradigms require for their efficient application to varying degrees the support by modern com­puter equipment and that in all three cases one of the most serious bottlenecks is to input data, information, or knowledge into the computer system before it can be pro­cessed efficiently. Hence some important areas, e.g. machine learning, were neglected at the beginning. In recent years this negligence has partly been recognized and cured. In Europe, for instance, two years ago a European Network of Excellence was started with the name COIL (Computational Intelligence and Learning) which includes, apart from the above mentioned three areas, also machine learning. The positive effects of this expansion are already becoming obvious and it is to be expected that other areas, for instance, from artificial intelligence will follow.

The most recent step in this direction is the inclusion of case based reasoning (CBR). This is certainly a well established area in AI, but so far it has not been a focus of soft computing. It deserves, therefore, high appreciation that a book is published that focuses on the interrelationship between CBR and the methods of what has so far been known as soft computing or computational intelligence. The publication of such a book is certainly a formidable task, which could at least at the present time only be solved by convincing the leading experts in the relevant areas to contribute to this volume. The editors of the book have to be congratulated that they have achieved exactly this!

In addition they have created a book that, even though very much at the forefront of scientific progress, is structured and written almost like a textbook. After a tutorial introduction into CBR, three chapters are dedicated to the application of fuzzy sets, neural nets and genetic algorithms to CBR, followed by a chapter that merges three areas (fuzzy sets, neural nets and CBR) and finally the book is concluded by three chapters which show the application of hybrid approaches to interesting and impor­tant real problems. This is certainly a book that will be of high interest to scientists as well as to practitioners and that presumably will become a classic and an often cited source in future publications in this area. I can only hope that very many readers will take advantage of this· unique opportunity and I can again congratulate the editprs of this book and the authors of the contributions in it for their excellent and professional work.

Aachen, March 2000 H.-J. Zimmermann

Preface

There has recently been a spurt of activity to integrate different computing paradigms such as fuzzy set theory, neural networks, genetic algorithms, and rough set theory, for generating more efficient hybrid systems that can be classified as soft computing methodologies. Here the individual tool acts synergetically, not competitively, for en­hancing the application domain of each other. The purpose is to provide flexible infor­mation processing systems that can exploit the tolerance for imprecision, uncertainty, approximate reasoning, and partial truth in order to achieve tractability, robustness, low solution cost and close resemblance with human like decision making.

Neuro-fuzzy computing, capturing the merits of fuzzy set theory and artificial neu­ral networks, constitutes one of the best-known visible hybridizations encompassed in soft computing. This integration promises to provide, to a great extent, more in­telligent systems (in terms of parallelism, fault tolerance, adaptivity, and uncertainty management) to handle real life ambiguous recognition/decision-making problems.

Case based reasoning (CBR) may be defined as a model of reasoning that incor­porates problem solving, understanding and learning, and integrates all of them with memory processes. It involves adapting old solutions to meet new demands, using old cases to explain new situations or to justify new solutions, and reasoning from precedents to interpret a new situation. Cases are nothing but some typical situations, already experienced by the system. A case may be defined as a contextualized piece of knowledge representing an experience that teaches a lesson fundamental to acnieving the goals of the system. The system learns as a by-product of its reasoning activity. It becomes more efficient and more competent as a result of storing the experience of the system and referring to them in later reasoning.

The case based system, in contrast to the traditional knowledge-based system, op­erates through a process of remembering one or a small set of concrete instances or cases and basing decisions on comparisons between the new situation and the old ones. Systems based on this concept are finding widespread applications in problems like medical diagnosis and law interpretation where the knowledge available is incomplete and/or evidence is sparse.

Research articles integrating CBR with a soft computing framework for develop­ing efficient methodologies and algorithms for various real life decision-making ap­plications have started to appear. Design of efficient knowledge based networks in the said paradigm is also being attempted. Here soft computing tools become effective in tasks like extracting cases from ambiguous situations and handling uncertainties, adapting new cases, retrieving old cases, discarding faulty cases, finding similarity between cases, maintaining an optimal size of case bases, and in approximate reason­ing for justifying a decision. This makes the system robust and flexible in its activity.

The present volume is aimed at providing a collection of fifteen articles, including a tutorial one, containing new material describing, in a unified way, the basic con­cepts and characteristic features of these theories and integrations, with recent devel­opments and significant applications. These articles, written by different experts over the world, demonstrate the various ways this integration can be made for designing

x Soft Computing in Case Based Reasoning

methodologies, algorithms and knowledge based networks for handling real life am­biguous situations efficiently. The first chapter by J. Main, T.S. Dillon and S.C.K. Shiu provides a tutorial introduction to CBR along with its basic features and an evaluation of its strengths and weaknesses in its current symbolic form, followed by the rele­vance of soft computing. Various application domains are also reviewed. This chapter is included for the convenience of understanding the theory of CBR and hence the remaining chapters of the book.

Chapters 2 and 3 focus on the role of fuzzy set theory, alone, in enhancing the capabilities of CBR systems. One may note that fuzzy logic, which is mainly con­cerned with providing algorithms for dealing with imprecision, uncertainty and ap­proximate reasoning, is the one primarily responsible for the emergence of the theory of soft computing. In Chapter 2, the authors H.-D. Burkhard and M.M. Richter have described the notions of similarity in the discipline of CBR and fuzzy set theory. Both local and global measures of similarity are discussed. It is shown how the concepts of CBR can be interpreted in terms of fuzzy set technology. A comparison of the reason­ing processes in CBR and fuzzy control is provided. E. Hiillermeier, D. Dubois and H. Prade have proposed, in Chapter 3, case based inferencing methodologies using fuzzy rules within the framework of approximate reasoning. Gradual inference rules and certainty rules are described, in this regard, as the basic model of inferencing, using the notion of fuzzy sets. Cases are viewed as individual information sources and case based inferencing is considered as a parallel combination of such information sources. A technique for rating the cases with respect to their contributions to the prediction of outcomes is provided.

The next two chapters deal with the hybridization of artificial neural networks which are reputed, as a soft computing tool, to provide the machinery for learning and curve fitting through its generic characteristics like adaptivity, massive parallelism, robustness and optimality. M. Malek has presented in Chapter 4 several hybrid ap­proaches for integrating neural networks and CBR. Various modes of integration with two degrees, namely, loose and tight coupling, are discussed. These are followed by a detailed description of a tightly coupled hybrid neuro CBR model, called ProBIS, for case retrieval adaptation and memory organization. Its effectiveness is demonstrated for medical diagnosis, control and classification problems. The task of classification is also addressed in Chapter 5 where C.-K. Shin and S.C. Park have described a hybrid system of neural networks and memory based reasoning for regression and classifica­tion. It incorporates a feature weighting mechanism which plays an important role in the underlying principle. This hybrid approach has synergetic advantages other than giving example based explanations together with prediction values of neural networks. Extensive experimental results are provided. The system is applicable to various re­search fields such as hybrid learning strategy, feature selection, feature weighting in CBR and rule extraction.

Chapters 6-8 concern the integration of CBR with another biologically inspired tool, namely genetic algorithms (GAs), along with some important illustrations. GAs are randomized search and optimization techniques guided by the principles of evo­lution and natural genetics. They are efficient, adaptive and robust search processes, producing near optimal solutions and having a large amount of implicit parallelism. W. Dubitzky and F.J. Azuaje have presented in Chapter 6 two approaches to (intro-

Preface xi

spective) learning case retrieval knowledge structures, namely learning of global and local feature weights using an evolutionary computing approach, and learning of use­ful case groupings using a growing cell structure neural model. The former method uses feedback of systems performance during the learning process, while the latter does not. The authors also provide a brief description of four general case retrieval models, namely computational, representational, partitioning and model based, and the basic features of GAs for the convenience of readers. The effectiveness of the methodologies is demonstrated on medical data for predicting coronary heart disease. In Chapter 7, a computational hybrid architecture for creative reasoning is described by A. Cardoso, E. Costa, P. Machado, EC. Pereira and P. Gomes. Here, CBR explores the previous knowledge and provides a long term memory, while the evolutionary computation complements with its adaptive ability. The cross-contribution, therefore, benefits creative reasoning. An example is shown demonstrating how a case library of images and a retrieval metric can be coupled with an evolutionary system in this regard. Chapter 8 describes a system called Teacher (Technique for the Automated Creation of Heuristics) for learning and generating heuristics used in problem solv­ing. Here B. W. Wah and A. Ieumwananonthachai employ a genetics based machine learning approach. The design process is divided into four phases, namely classifica­tion, learning, performance verification and performance generalization. The objective is to determine, under resource constraints, improved heuristics methods as compared to existing ones. Experimental results on heuristics learned for problems related to circuit testing are demonstrated.

So far we have discussed the role of fuzzy sets, artificial neural networks and genetic algorithms, as an individual soft computing tool, in case based systems. The next four chapters highlight the importance of integration of two of these tools, namely fuzzy sets and neural networks, in a similar context. One may note that neural fuzzy integration or neuro-fuzzy computing is proved to be the most visible hybrid soft com­puting tool with generic and application specific merits available so far. The literature in this area is also considerably rich compared to other integrated soft computing paradigms. Of the four articles considered here, Chapter 9 discusses the role of fuzzy neurons in handling uncertainties in input data, while Chapters 10, 11 and 12, which are from the groups of the editors, deal with the problems of case selection/deletion and case retrieval.

A case based decision support system is designed in Chapter 9 by Z.-Q. Liu. Here, the network model uses two types of fuzzy neurons, namely fuzzy AND neuron and fuzzy OR neuron. The system is capable of handling uncertainties in linguistic inputs and contractory data. Its effectiveness is demonstrated rigorously in many domains including telecommunications. The case selection method described by R.K. De and S.K. Pal (Chapter 10) includes a design procedure for a layered network for selec­tion of ambiguous cases through growing and pruning of nodes. The notion of fuzzy similarity between cases is used. The effectiveness of the network is demonstrated on speech recognition and medical diagnosis problems. The performance is validated by a I-NN classifier with the selected cases as input. The authors have also described, in brief, a few existing case based systems for ready reference. Chapter 11 by S.C.K. Shiu, X.Z. Wang and D.S. Yeung provides a method of selecting the representative cases for maintaining the size of a case base without losing much of its competence.

xii Soft Computing in Case Based Reasoning

The algorithm involves computation of feature importance, enhancement of classifi­cation in the transformed feature space, and computation of case density. The features are demonstrated with a real life glass classification example. In Chapter 12, J. Main and T.S. Dillon mainly focus on showing how to use fuzzy feature vectors and neural networks for improving the indexing and retrieval steps in a case based system. The feasibility of such a system is illustrated on the problem of fashion footwear design.

The remaining three chapters deal with some more real life significant applications under the aforesaid different modes of integration. In Chapter 13, J.M. Corchado and B. Lees have presented a connectionist approach for oceanographic forecasting using the notion of CBR. The system combines the ability of a CBR system for selecting previous similar situations and the generalizing ability of artificial neural networks to guide its adaptation stage. The system is able to forecast the thermal structure of water ahead of an ongoing vessel. Here the CBR method is used to select a number of cases (from a large case base), and the neural network produces final forecasts in real time based on these selected cases. The system has been successfully tested in the Atlantic Ocean. O. Stahl in Chapter 14 has dealt with an interesting application in space mis­sion by designing a fuzzy case based system (called CREW) which can predict the key psychological factors, e.g., stress, morale, teamwork, of the crew members in an astronaut team over time. The work was carried out at NASA in order to select crew members for the international space station. Here the notion of CBR is used over a rule based approach as the latter would need to acquire a large number of formal rules ~apable of computing predictive results for all the combinations of 12 psychological factors and eight event types. The theory of fuzzy sets helps in encoding various con­ditions of adapting rules in terms of linguistic variables, e.g., low, medium and high, for matching similar old cases to a new one.

Chapter 15 demonstrates how the principle of CBR and fuzzy sets together can be applied to the areas of medical equipment diagnosis, plastics color matching, resi­dential property valuation and aircraft engine monitoring, for the purpose of selecting cases. Here B. Cheetham, P. Cuddihy and K. Goebel use the theory of fuzzy sets to find similarity between cases. For medical equipment diagnosis and residential property valuation, membership functions characterizing fuzzy sets provide greater selection accuracy through their noise tolerance. In the case of a plastic color matching system, the use of multiple selections allows detection of potential problems during the case selection phase. Incorporation of adaptive fuzzy clustering technique for fault classi­fication in aircraft engines enhances the ability to deal with extremely noisy sensors and to adjust to unpredictable slow drift.

This comprehensive collection provides a cross-sectional view of the research work that is currently being carried out over different parts of the world applying CBR in a soft computing framework. The book, which is unique in its character, will be useful to graduate students and researchers in computer science, electrical engi­neering, system science, and information technology not only as a reference book, but also a textbook for some parts of the curriculum. Researchers and practitioners in in­dustry and R&D laboratories working in the fields of system design, control, pattern recognition, data mining and vision will also benefit.

We take tliis opportunity to thank all the contributors for agreeing to write for the book. We owe a vote of thanks to Ms. Karen Barker of Springer-Verlag London for

Preface xiii

her initiative and encouragement. The technical/software assistance provided by Dr. Rajat K. De, Mr. Pabitra Mitra, Mr. Suman K. Mitra and Dr. Simon Shiu, and the sec­retarial help by Mr. I. Dutta and Mr. S. Das during the preparation of the book, are gratefully acknowledged. Particular mention must be made of Pabitra for his sincerity, enthusiasm and dedicated effort, without which it would have been difficult to com­plete the project in time. The work was initiated when Prof. S.K. Pal was working as a visiting professor in the Department of Computing, Hong Kong Polytechnic, Hong Kong, during April-June 1999.

February 2000 S.K. Pal T.S. Dillon D.S. Yeung

Contents

List of Contributors ............................................... xvii

1. A Thtorial on Case Based Reasoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Julie Main, Tharam S. Dillon and Simon C.K. Shiu

Fuzzy Sets

2. On the Notion of Similarity in Case Based Reasoning and Fuzzy Theory 29 Hans-Dieter Burkhard and Michael M. Richter

3. Formalizing Case Based Inference Using Fuzzy Rules. . . . . . . . . . . . . . .. 47 Eyke Hiillermeier, Didier Dubois and Henri Prade

Artificial Neural Networks

4. Hybrid Approaches for Integrating Neural Networks and Case Based Reasoning: From Loosely Coupled to Tightly Coupled Models . . . . . . .. 73

Maria Malek

5. Towards Integration of Memory Based Learning and Neural Networks. 95 Chung-Kwan Shin and Sang Chan Park

Genetic Algorithms

6. A Genetic Algorithm and Growing Cell Structure Approach to Learn-ing Case Retrieval Structures .................................... 115 Werner Dubitdcy and Francisco J. Azuaje

7. An Architecture for Hybrid Creative Reasoning .................... 147 Amilcar Cardoso, Ernesto Costa, Penousal Machado, Francisco C. Pereira and Paulo Gomes

8. Teacher: A Genetics Based System for Learning and Generalizing Heuristics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Benjamin W. Wah and Arthur leumwananonthachai

Neuro-Fuzzy Computing

9. Fuzzy Logic Based Neural Network for Case Based Reasoning ........ 213 Zhi-Qiang Liu

xvi Soft Computing in Case Based Reasoning

10. Case Based Systems: A NeullO-Fuzzy Method for Selecting Cases .... 241 Rajat K. De and Sankar K. Pal

11. Neuro-Fuzzy Approach for Maintaining Case Bases ................ 259 Simon C.K. Shiu, X.Z Wang and Daniel S. Yeung

12. A Neuro-Fuzzy Methodology for Case Retrieval and an Object­Oriented Case Schema for Structuring Case Bases and their Application to Fashion Footwear Design .......................... 275 Julie Main and Tharam S. Dillon

Applications

13. Adaptation of Cases for Case Based Forecasting with Neural Network Support ..................................................... 293 J.M. Corchado and B. Lees

14. Armchair Mission to Mars: Using Case Based Reasoning and Fuzzy Logic to Simulate a Time Series Model of Astronaut Crews .......... 321 Gerry Stahl

15. Applications of Soft CBR at General Electric ...................... 335 Bill Cheetham, Paul Cuddihy and Kai Goebel

About the Editors ... , .............................. , .............. 371

List of Contributors

• Francisco Azuaje Department of Computer Science Trinity College, University of Dublin Dublin 2 Ireland. E-mail: francisco.azuaje@cs.tcd.ie

• Hans-Dieter Burkhard Institut fur Informatik Humboldt Universitat Zu Berlin Unter den Linden 6 10099 Berlin Germany E-mail: hdb@informatik.hu-berlin.de

• Amflcar Cardoso Centro de Informaticae Sistemas da Universidade de Coimbra (CISUC) Pinhal de Marrocos 3030 Coimbra Portugal E-mail: amilcar@deLuc.pt

• Bill Cheetham GE Research and Developement Center Building K-l Room 5C21A 1 Research Circle Niskayuna, NY 12309 USA E-mail: cheetham@crd.ge.com

• Juan M. Corchado Artificial Intelligence Research Group Escuela Superior de Ingenieria Informatica University of Vigo Campus Universitario As Lagoas Edificio Politecnico Ourense 32004 Spain E-mail: corchado@eLuvigo.es

• Ernesto Costa Centro de Informaticae Sistemas da Universidade de Coimbra (CISUC) Pinhal de Marrocos 3030 Coimbra Portugal E-mail: ernesto@deLuc.pt

• Paul Cuddihy GE Research and Developement Center Building K-l Room 5C21A 1 Research Circle Niskayuna, NY 12309 USA E-mail: cuddihy@crd.ge.com

• RajatK. De Machine Intelligence Unit Indian Statistical Institute 203 B. T. Road Calcutta 700 035 India E-mail: rajat@isical.ac.in

• Tharam S. Dillon Department of Computer Science and Computer Engineering La Trobe University Bundoora, Victoria 3083 Australia E-mail: tharam@cs.latrobe.edu.au

• Werner Dubitzky German Cancer Research Centre Intelligent Bioinformatics Systems 1m Neuenheimer Feld 280 69120 Heidelberg Germany E-mail: W.Dubitzky@dkfz-heidelberg.de

xviii

• Didier Dubois Institut de Recherche en Informatique de Toulouse Universit Paul Sabatier 118, route de Narbonne 31062 Toulouse France E-mail: dubois@irit.fr

• Kai Goebel GE Research and Developement Center Building K-1 Room 5C21A 1 Research Circle Niskayuna, NY 12309 USA E-mail: goebel@crd.ge.com

• Paulo Gomes Centro de Informaticae Sistemas da Universidade de Coimbra (CISUC) Pinhal de Marrocos 3030 Coimbra Portugal E-mail: pgomes@deLuc.pt

• Eyke Hiillermeier Institut de Recherche en Informatique de Toulouse Universit Paul Sabatier 118, route de Narbonne 31062 Toulouse France E-mail: eyke@irit.fr

• Arthur Ieumwananonthachai NonStop Networking Division Tandem Computers Inc. 10501 N. Tantau Avenue, LOC201-02 Cupertino, CA 95014-0728 USA E-mail: ieumwananonthachai _arthur@tandem.com

Soft Computing in Case Based Reasoning

• BrianLees Applied Computational Intelligence Research Unit Department of Computing and Information Systems, University of Paisley Paisley PAl 2BE UK E-mail: lees-ci@paisley.ac.uk

• Zhi-Qiang Liu Computer Vision and Machine Intelligence Lab (CVMIL) Department of Computer Science University of Melbourne 221 Bouverie Street, Carlton Victoria 3053 Australia E-mail: zliu@cs.mu.oz.au

• PenousalMachado Centro de Informaticae Sistemas da Universidade de Coimbra (CISUC) Pinhal de Marrocos 3030 Coimbra Portugal E-mail: machado@dei.uc.pt

• Julie Main Department of Computer Science and Computer Engineering La Trobe University Bundoora, Victoria 3083 Australia E-mail:main@cs.latrobe.edu.au

• Maria Malek EISTI -Computer Science Department Ave du Parc, 95011 Cergy France E-mail:maria.malek@wanadoo.fr

List of Contributors

• Sankar K. Pal Machine Intelligence Unit Indian Statistical Institute 203 B. T. Road Calcutta 700 035 India E-mail: sankar@isical.ac.in

• Sang Chan Park Department of Industrial Engineering Korea Advanced Institute of Science and Technology Yusong, Taejon 305-701 South Korea E-mail: sangpark@major.kaist.ac.kr

• Francisco C. Pereira Centro de Informaticae Sistemas da Universidade de Coimbra (CISUC) Pinhal de Marrocos 3030 Coimbra Portugal E-mail: camara@deLuc.pt

• Henri Prade Institut de Recherche en Informatique de Toulouse Universit Paul Sabatier 118, route de Narbonne 31062 Toulouse France E-mail: prade@irit.fr

• Michael M. Richter Institut fur Informatik Humboldt Universitat Zu Berlin Unter den Linden 6 10099 Berlin Germany E-mail: richter@informatik.uni-kl.de

• Chung-Kwan Shin Department of Industrial Engineering KoreaAdvanced Institute of Science and Technology Yusong, Taejon 305-701

South Korea E-mail: ckshin@major.kaist.ac.kr

• Simon C.K. Shiu Department of Computing Hong Kong Polytechnic University Hung Hom, Kowloon Hong Kong

xix

E-mail: cskshiu@comp.polyu.edu.hk

• Gerry Stahl Center for Life Long Learning and Design Department of Computer Science and Institute of Cognitive Science University of Colorado, Boulder CO 80309-0430 USA E-mail: gerry. stahl @colorado.edu

• Benjamin W. Wah Electrical and Computer Engineering Coordinated Science Laboratory 446 Computer and Systems Research Laboratory, MC-228 1308 West Main Street Urbana, IL 61801 USA E-mail: b-wah@uiuc.edu

• X.Z. Wang Department of Computing Hong Kong Polytechnic University Hung Hom, Kowloon Hong Kong E-mail: csxzwang@comp.polyu.edu.hk

• Daniel S. Yeung Department of Computing Hong Kong Polytechnic University Hung Hom, Kowloon Hong Kong E-mail: csdaniel@comp.polyu.edu.hk