WISDOM
Web Intelligent Search based on DOMain ontologies
The huge amount of data and services available on the Web requires the development of systems able to overcome
the "information overloading" problem of traditional search engines. In particular, there is the need of developing
novel tools for the integration, the localization and the customizable fruition of informative resources, thus
allowing clients to "recharge" with interesting data.
The goal of the MIUR-funded WISDOM project is to develop intelligent techniques and tools, based on domain ontologies,
to perform effective and efficient information search on the Web. In particular, we aim at developing systems for
retrieving information both from data-intensive and unstructured sites/web pages, in an integrated and efficient way.
The project will be articulated into 3 synergic and complementary themes and will define a reference methodological
and functional architecture in order to ensure compatibility of the solutions offered by the three themes.
- The goal of Theme1 (Creation and extension of a domain ontology) is the study and development of solutions to
represent the semantics of the contents of Web sources.
- The goal of Theme2 (Emergent Semantics: Discovering semantic mappings among domain ontologies) is the
study and development of techniques and tools to support identification, discovery and storage of semantic
mappings among domain ontologies. The investigated semantic mappings will support the rewriting of a query
with respect to more ontologies and will be based on language semantics, lexical chains and logic inferences.
- The goal of Theme3 (Query processing) is the development of techniques for searching information by
exploiting the semantic infrastructure developed within Themes 2 and 3. Efficient and effective query
processing mechanisms, considering data/sites hetereogeneity and constraints imposed by the distributed
environment, will be developed. In particular, these techniques will rely on sources characterization to
individuate useful sources, solve rewriting problems and integrate results from different sources.
Partners
Links