Hierarchical knowledge oriented specification for information integration Madalina Croitoru Department of Computing Science, University of Aberdeen Aberdeen, UK Ernesto Compatangelo Department of Computing Science, University of Aberdeen Aberdeen, UK 30th June 2005 Abstract We present a novel methodology for manipulating sources in a knowledge integration scenario. First we define and exploit an appropriate data model – Knowledge Oriented Specification – for representing and query- ing the data sources without having to align their background knowledge. Second we propose a structured knowledge representation formalism – Layered Conceptual Graphs – which present the data at different levels of detail. We explain how the two formalisms can be used together to provide a hierarchical approach to integration. In this way, the user – according to their interest in the integrated topic – will be presented with different views on the combined knowledge. 1 Introduction Information integration is a broad notion that encompasses the combination of different kinds of distributed, heterogeneous sources such as databases and web sites. The integration process generates a ’unified’ view of two or more indi- vidual sources; this view can be either virtual (if the sources remain physically distinct) or concrete (if the sources are physically merged) [24]. The user of an information integration system is then able to query the unified view resulting from the integration process, getting results that could not be retrieved from any of the individual sources if independently considered. One of the main problems in the integration process is the semantic mis- match between sources. Approaches attempting to solve these problems can be grouped into two different categories: multi-agent systems and global in- formation systems (based on databases or based on ontologies). Intelligent In- formation Integration has been a popular application for multi agent systems. 1