With the explosion in the availability of information online, users are finding it increasingly difficult to track down the specific material that they require. Users are therefore becoming increasingly dependent on intelligent services to provide information that is dynamically selected and presented according to their preferences. However, development of these personalized services is not trivial. Significant effort is required in terms of engineering the underlying knowledge that is used by a service to determine which information might be relevant to a particular user and how to present it. A Toolkit that reduces the complexity involved in the creation and maintenance of such services is discussed.

Good communication is vital in health care, both among health care professionals, and between health care professionals and their patients. And well-written documents, describing and/or explaining the information in structured databases may be easier to comprehend, more edifying, and even more convincing than the structured data, even when presented in tabular or graphic form. Documents may be automatically generated from structured data, using techniques from the field of natural language generation. These techniques are concerned with how the content, organization and language used in a document can be dynamically selected, depending on the audience and context. They have been used to generate health education materials, explanations and critiques in decision support systems, and medical reports and progress notes.

This paper describes a current research project investigating belief revision in intelligent systems by modelling the librarian in interaction with a literature-seeking user. The work is designed both to test a theory of agent behaviour based on belief revision proposed by Galliers, and to evaluate a model of the librarian developed by Belkin, Brooks and Daniels, through computational implementation. Agent communication is seen as motivated by and motivating belief changes, where belief revision is determined by coherence, combining endorsement, connectivity and conservatism. The librarian is viewed as a distributed expert system with many individual specialised functions operating in particular belief domains. The paper describes our first implementation of the belief revision mechanism and of a very primitive librarian, designed to test the basic viability of our ideas and to allow us to explore different forms of the distributed system architecture.

Explanations are important in many areas of human-computer interaction. In help systems, tutoring systems and within expert system, lengthy explanations of some topic or justifications of some reasoning process may be required. If a long explanation is given, there is a good chance that at some point the user will ‘lose track’, and fail to grasp the main content of the explanation. There has therefore been recent emphasis on generating explanations and textual descriptions that are tailored to the knowledge and goals of the particular user. However, there is no guarantee that such a model will be accurate. By allowing interactions with the user within the explanation this no longer becomes crucial. Then, if users are confused in the middle of an explanation they can interrupt and seek clarification, and the system may provide explicit checks on the user's understanding. Therefore this paper presents an approach to explanation generation based on the assumption that explanations must both use and track a model of what the user knows, and also involve interactions with the user. The framework is based on sociolinguistic studies of human-human interaction as well as artificial intelligence work on explanation, text planning, tutoring and user modelling. It has been implemented and used for generating tutorial explanatory dialogues in electronics.