Temple University

Research Title: 
Removing Semantically Incorrect Results from Information Extraction from Text
Collaborator Type: 
Academia

Our group's research is in the areas of natural language processing, data mining, and machine learning. We expect ResearchCyc to be useful for our research efforts on information extraction. We are interested in experimenting with using it to help weed out incorrect extractions that our current systems discover in Web text, by eliminating extractions that contradict common-sense knowledge in the Cyc database. Second, we intend to investigate techniques for disambiguating our extracted terms by connecting them to symbols in the Cyc ontology.