Modelling semantic grid knowledge embedded in documents

The growing success of grid technologies inside scientific communities has produced an increasing need for the development of tools and methodologies able to support knowledge sharing and handling among people, built upon the grid. This "semantic" infrastructure is becoming to be referred as semantic grid. In this paper we propose an original approach to the development of a system for the creation of the knowledge layer of the semantic grid, that is, the layer which carries the informative content that the community shares. Using well-assessed natural language processing and machine learning methodologies and techniques, our goal is to acquire and organize the information stored in the grid, where this information is supposed to be represented in unstructured documents. Our intent is to extract and shape knowledge in syntactic patterns and organize them into a hierarchy of relational concepts, whose goal is to improve the process of knowledge retrieval and maintenance.