One of the key factors for the acceptance of expert systems in real-world domains is the ability to explain their reasoning (Buchanan & Shortliffe, 1984; Henrion & Druzdzel, 1990). This paper describes the basic properties that characterise explanation methods and reviews the methods developed to date for explanation in Bayesian networks.

The use of methodologies in software and knowledge engineering is very extensive due to their important advantages. In the case of the development of ontologies, until now, several methodological proposals have been presented for building ontologies. Some of these methodologies are designed for building ontologies from scratch or reusing other ontologies without modifying them, concretely, the following cases can be mentioned: the Cyc methodology, the approach proposed by Uschold and King, Grüninger and Fox's methodology, the KACTUS methodology, METHONTOLOGY and the SENSUS methodology. There is even a proposal for re-engineering ontologies, and several proposals for collaborative construction of ontologies.

In this article, we describe the methodologies and check their degree of maturity, contrasting them with respect to the IEEE standard for software development. Before this, we justify to what extent this standard can be used. A conclusion to this study is that there is no completely mature methodological proposal for building ontologies, since there are some important activities and techniques that are missing in all these methodologies. However, all the methodologies do not have the same degree of maturity. In fact, METHONTOLOGY is a very mature methodology. The other conclusion of this article is that, although work to unify proposals can be interesting, maybe several approaches should coexist.

This survey introduces existing approaches to Agent Communication Languages (ACLs) and particularly Conversation Policies (CPs) which can be viewed as general constraints on the sequence of semantically coherent messages leading to a goal. Then limitations of these CPs are discussed in detail, particularly limitations on flexibility and specification. Finally, ACLs are viewed from the dialectic point of view, and some approaches are introduced in this context: some focusing on commitment-based protocols and others on dialogue-game-based protocols.

1 Believable interactive embodied agents

Among the goals of research on autonomous agents one important aim is to build believable interactive embodied agents that are apt to application to friendly interfaces in e-commerce, tourist and service query systems, entertainment (e.g. synthetic actors) and education (pedagogical agents, agents for help and instruction to the hearing impaired).

1 Opening

PKDD 2001, the 5th European Conference on Principles of Knowledge Discovery in Databases (PKDD), was held in Freiburg, Baden-Württemberg, Germany, this year (Monday 3 to Thursday 7 September), and co-located with the 12th European Conference on Machine Learning (ECML 2001). The proceedings comprised two volumes, one for PKDD (De Raedt & Siebes, 2001) and one for ECML (De Raedt & Flach, 2001); and form part of the Springer Lecture Notes on Artificial Intelligence (LNAI) series. The conference was held in the University buildings in the centre of the old town. Freiburg and the surrounding area were for many years part of the Austro-Hungarian empire and thus the university was described to us as being one of the oldest Austrian Universities.

Bayesian Networks and Decision Graphs by Finn V. Jensen, Springer Verlag, £64.95, ISBN 0-387-95259-4