Ontologies are the subject of much interest these days, because they are seen as the cornerstone of the semantic web, data interoperability and many knowledge intensive natural language and artificial intelligence applications. Today, most of the ontology construction is done manually, and is a very time consuming process. Automating this process is of great interest.
This talk presents the research and design of a Knowledge Acquisition from Text (KAT) system that attempts to extract concepts and semantic relations straight from text, and classify these into ontological hierarchies.
Extensive natural language processing capabilities are needed for this.
The domain - specific ontologies produced by KAT are further integrated with other exiting higher level ontologies.
The talk will also compare this research with other efforts on building ontologies.