×

Advances in knowledge acquisition. 9th European knowledge acquisition workshop, EKAW ’96, Nottingham, GB, May 14-17, 1996. Proceedings. (English) Zbl 0864.68084

Lecture Notes in Computer Science 1076. Lecture Notes in Artificial Intelligence. Berlin: Springer-Verlag. xii, 369 p. (1996).
This volume contains the proceedings of the ninth European Knowledge Acquisition Workshop (EKAW’96), held at Nottingham, England, May 1996. The work reported in this volume (altogether 23 articles) can be divided into five major categories, viz. theoretical and general issues of knowledge acquisition, the elicitation of knowledge from textual and other sources, data mining, group knowledge elicitation, and planning. The first group of theoretical papers is concerned with fundamental questions about the forms problem solving methods should take. These papers deal with formal languages for selecting appropriate problem solving methods based on underlying assumptions, efficiency aspects of properly chosen problem solving methods and, more general, the knowledge acquisition process itself, and libraries of reusable problem solving methods. The second group of theoretical papers is mainly concerned with questions of language and ontology. Methods for comparing languages for expressing problem solving methods are discussed which focus on the goals they try to achieve, formal methods are introduced for expressing knowledge in a diagnostic reasoner, and methodologies are proposed for the construction of ontologies for technical domains. The second major topic, the extraction of knowledge from rich but ill-structured sources, is dealt with in papers which discuss the application of statistical clustering methods for text analysis in order to determine significant terms for knowledge acquisition, propose a quality-based terminological reasoning methodology for dealing with the significance of concept hypotheses as derived by a text understanding system, or deal with the extraction of knowledge from large texts by identifying specific conceptual relations in the text (such as definition and exemplification).
As far as knowledge acquisition from non-textual sources is concerned, a case study dealing with the exploitation of the World Wide Web and a model for the acquisition of knowledge from diagrams are reported. Data mining from large databases using machine learning methodologies is focused on in contributions which take an evaluative look at current methodologies, e.g., considering the quality, in terms of understandability, of automatically induced theories or the user-friendliness of conceptualizations provided by machine learning techniques. Other papers present cognitively adequate tools for incrementally generating concept hierarchies from a multi-perspective point of view, propose a redescription methodology by which the discrepancy between the format of databases used for data mining and the formats needed by machine learning methods can be resolved and consider higher conceptual abstractions to ease the knowledge acquisition process.
Knowledge acquisition with groups of experts poses additional problems which are considered in papers that investigate the use of a shared task model as a basis for interactions between expert users and knowledge engineers, or introduce methods for constructing a shared knowledge base, combining Warfield’s brainwriting technique of amassing viewpoints of particular issues and a decision support system for constructing a shared memory. The final group of papers is concerned with knowledge acquisition in planning domains. Industrial applications are described to elicit the methods used by schedulers to amend schedules in the dynamic environment of automobile manufacturing using formalized methods and software support for schedule repair; also, tools for detecting certain classes of error in and tools for interactive debugging of planning knowledge bases are introduced, the use of the CommonKADS library and methodology is illustrated in the service recovery planning domain, and knowledge-level modelling techniques are applied to a search and rescue planing domain.

MSC:

68T05 Learning and adaptive systems in artificial intelligence
68-06 Proceedings, conferences, collections, etc. pertaining to computer science
68T35 Theory of languages and software systems (knowledge-based systems, expert systems, etc.) for artificial intelligence

Software:

CommonKADS
PDFBibTeX XMLCite