Researchers in AI are deeply involved in Semantic Web development, working on such topics as standardized ontologies, formal foundations for ontologies, new representation languages and the adaptation of old languages to the web.

At the same time, researchers in computational linguistics are developing means to adequately represent linguistically annotated data, with the goal of developing formats and standards that will eventually enable full exploitation of the information represented. They are increasingly turning toward resources developed within the XML framework such as the Resource Definition Framework (RDF) to model the information in ways that will allow for maximal flexibility and extensibility. This demands, in turn, development of abstract models that capture the properties of linguistic annotations at various levels of specificity, and development of ontologies to represent them. The need to develop and standardize representation formats for linguistic data and its annotations has grown to the point where a new working group has been formed within the International Standards Organization (ISO) to oversee this activity. However, much of this activity is going on with only superficial knowledge of developments in the framework of the Semantic Web and the potential for accessing and exploiting information that it is intended to eventually allow.

This workshop brings together researchers in AI who are working on the Semantic Web and those involved in the development of standards for linguistic annotation, to enable an exchange of information and ideas. This is a critical point at which to bring together these two groups, who typically have little interaction. Those involved in developing language resources need to gain a deeper understanding of the potential of and requirements for the Semantic Web and standardized ontologies, and AI researchers, who are working on a general model, will gain insight by considering an application of their work to actual content and, more generally, by considering the needs for a specific domain that requires complex representation mechanisms and sophisticated means to exploit the information. 

Paul Buitelaar, DFKI, Saarbrucken, Germany
Nicoletta Calzolari, ILC-CNR, Italy
Christiane Fellbaum, Princeton, USA
Aldo Gangemi, ITBM-CNR, Italy
Nicola Guarino, LADSEB-CNR, Italy
Graeme Hirst, University of Toronto, Canada
Atanas Kiryakov, SIRMA Ontotext Lab, Bulgaria
Sergei Nirenburg, New Mexico State University, USA
James Pustejovsky, Brandeis University, USA
Laurent Romary, LORIA/INRIA, France