Intelligent Information Systems 2002

Citation

(2003), "Intelligent Information Systems 2002", Kybernetes, Vol. 32 No. 7/8. https://doi.org/10.1108/k.2003.06732gae.005

Intelligent Information Systems 2002

Intelligent Information Systems 2002: Proceedings of the IIS 2002 Symposium, Sopot, Poland, 3-6 June 2002

Mieczysøaw A. Køopotek, Søawomir T. Wierzchon´ andMaciej Michalewicz (Eds)Physica-Verlag (A Springer-Verlag Company)Heidelberg2002ISBN 3-7908-1509-8, ISSN 1615-3871x + 468 pp.Paperback, £46.00Advances in Soft Computing SeriesReview DOI 10.1108/03684920310483306

This is the proceedings of a symposium held in a resort location on the Gulf of Danzig. It is the 11th in the general area of artificial intelligence organised by the Institute of Computer Science of the Polish Academy of Sciences, the first being in 1992. The "Advances in soft computing" series is designed to allow short publication time and worldwide distribution, and the former has undoubtedly been achieved here.

The 49 papers in the present volume are grouped under five headings.

1. 1.

   Decision Trees and Other Classifier Systems (14 papers),

2. 2.

   Neural Network and Biologically Motivated Systems (seven papers),

3. 3.

   Clustering Methods (seven papers),

4. 4.

   Handling Imprecision and Uncertainty (11 papers), and

5. 5.

   Deductive, Distributed and Agent-based Systems (ten papers).

The great majority of the papers are by workers in Poland, with four having joint Polish and American authorship. One paper shows only American affiliations and another only British. There is one paper from Belarus and another with joint authors from Britain, France and Ukraine. In one of the papers with joint American authorship, the author having an American affiliation also has a Polish one and is R.S. Michalski, a well-established authority on machine learning. Two of the papers having Polish authorship have authors from a Polish-Japanese Institute of Information Technology in Warsaw.

The subject matter obviously ranges widely. In their Preface, the editors express their satisfaction that, since the previous events, there is an increase in the diversity of practical applications. The first paper in the book has an important medical application, since it refers to the use of data mining techniques to help diagnose melanoma. The paper by R.S. Michalski as a joint author, under heading (1), is about the modelling of computer user behaviour. Other papers that mention specific applications refer to e-commerce and deductions from financial data, and another is on the extraction of information from audio signals, where the information includes features relevant to speaker and mood identification.

The seven papers under heading (2) show a fine diversity of approach, since the first refers to a computing technique inspired by an ant colony, the next two are about neural nets, but treating them in quite different ways, and the next is about an artificial immune system. These are followed by two papers dealing with genetic algorithms. The final paper under the heading is about an approach using quantitative property structure relations (QSPRs) to infer physical properties of chemical compounds, in this case, their boiling points.

The first paper under heading (3) is the one, whose authors have British affiliations and is on the intriguing topic of the use of AI techniques for the synthetic generation of crowd scenes in TV or film productions. Another paper in this section discusses double clustering, which is essentially the two-stage process in which ordinary clustering is performed first and then the centres of the clusters are themselves clustered. This paper has a medical application, since the use of the method is illustrated with reference to a study relating cardiac arrhythmia to meteorological data.

The papers under heading (4) range widely in topic and refer to special algebras and logics and developments in knowledge representation, data mining, reinforcement learning and Bayesian belief networks. Those under heading (5) also range widely and two of them make specific reference to the Internet, one of them proposing an expert system applied to intelligent web search that should permit powerful worldwide market analysis.

This review of some contents of the symposium is inevitably patchy and biased towards the contributions mentioning immediate practical applications, though the more abstract treatments may well be even more significant. The symposium was undoubtedly, productive of many new ideas and methods and these proceedings are a valuable account of them.

Alex M. Andrew