Topic Extraction and Extension to Support Concept Mapping (pdf )

David B. Leake, Ana Maguitman, and Thomas Reichherzer. Proceedings of the Sixteenth International Florida Artificial Intelligence Research Society Conference (FLAIRS-2003), AAAI Press, 2003, pp. 325-329


Successful knowledge management may depend not only on knowledge capture, but on knowledge construction---on formulating new and useful knowledge that was not previously available. Electronic concept mapping tools are a promising method for supporting knowledge capture and construction, but users may find it difficult to determine the right knowledge to include. Consequently, knowledge-based methods for suggesting relevant information are desirable for supporting the knowledge modeling process. We are developing methods to aid concept mapping by suggesting relevant information to compare, contrast, and possibly include in knowledge models represented as concept maps. This paper presents two specific methods we are developing for this task, both of which automatically identify topics related to a concept map in order to guide the retrieval of related information. The first, DISCERNER, automatically organizes concept map libraries into a hierarchical structure of topic categories and subcategories that are used as indices for efficient access to relevant stored concept maps. The second, EXTENDER, characterizes the topics of concept maps under construction, applies clustering techniques to the resulting information, and performs incremental web-mining for new but related, topics. It suggests these topics as potential areas for extending the existing concept map or to include in new maps to increase current knowledge coverage.

See for additional publications in the Artificial Intelligence/Cognitive Science report and reprint archive maintained by David Leake.