Roxana Gîrju and Dan Moldovan, Southern Methodist University, USA
Questions can be classified based on their degree of difficulty. As the level of difficulty increases, question answering systems need to rely on richer semantic ontoiogies and larger knowledge bases. This paper is concerned with questions whose answers are spread across several documents and thus, require answer fusion. To find such answers, the system needs to develop domain specific ontologies. A method is presented for on-line acquisition of ontological information from the document collection.