WP1 is concerned with the semi-automatic generation of ontologies. Tasks are divided into the three main groups: (1) data preparation, (2) generation of ontologies, and (3) advanced topics. Initial work will focus on various issues regarding data preparation for later ontology construction with some strong relationships to the tasks in WP2. The two key deliverables appear in the second phase (1) on analytical evaluation of ontologies (horizontal task between WP1, WP2, and WP3) and (2) construction of the ontologies using combination of combinatorial optimization techniques and constrained hierarchical clustering (in cooperation with WP3). During the latter part of 18 month period some more advanced topics will be addressed such as “stream ontologies” (conceptualizing stream of data), “simultaneous ontologies” (modeling different views to the same data) and usage of “upper level ontologies” for ontology construction (relation to WP2). By the end of 18 month period, there will be specifications and preliminary prototypes for such systems.

WP2 will use human language technology to semi-automatically extract metadata. One task will extract metadata against a pre-defined ontology. During the 18 month period, the prototype for ontology creation with hierarchical clustering will be used to generate such ontologies. Another task will look at the use of a controlled language to simplify the task of metadata extraction. This is a particularly useful approach for small organizations and individuals wishing to make data available on the Semantic Web, as an alternative to authoring tools. Another task will look at the rhetorical structure in data, as an aid to metadata extraction. All the preceding work in this workpackage will be integrated into metadata annotation tools. Finally there will be a task to quantitatively evaluate the metadata extraction tools generated in this workpackage.

WP3 looks ontology and metadata management, and specifically evolution. Building on a study of existing ontology and metadata management facilities, the workpackage will first develop ontology evolution techniques and then techniques for metadata evolution. Finally, these two forms of evolution will be integrated. As ontology evolve inconsistencies develop, and hence this workpackage will include work on reasoning with inconsistencies. WP3 will be aligned and connected to the technologies developed in WP1, WP2 and WP4. Last, but not least, WP3 will be integrated in other existing technologies in WP6.

WP4 will develop software to undertake ontology mediation. During the first 18 months of the project will be reports on the state of the art in ontology mediation; on the establishment if an ontology mediation patterns library; on ontology mediation management; and on ontology mediation as a web service. There will also be specific guidelines for ontology mediation in each of the three case studies.

WP5 has seven deliverables. A survey of knowledge access tools for the semantic web will be the first deliverable. A user profiling module will be developed which tracks behaviour when accessing knowledge repositories. A “knowledge generation” module will be developed using text summarization and other HLT technology to deliver knowledge at the appropriate level of granularity. Initial prototypes of tools for searching, sharing and visualsing semantically annotated knowledge will be developed. Finally, a knowledge repurposing module will account of the recipient device in order to deliver knowledge to knowledge workers “any time, any place”.

WP6 will take the work of the preceding workpackages and integrate the various software modules into a comprehensive suite of tools.