Use of semantic information in a SyntacticDependency parser

Abstract

This paper describes some extensions of the parser that has been used in the Evalita 2011 contest. The paper does not focus on the actual parser, since it was already used in Evalita 2009: describing it would be a duplication of the description already given there. On the contrary, the paper addresses two extensions that have been adopted in a more recent version of the parser. The reason why this version was not used in Evalita is that it is based on a domain ontology not available for open domain texts and that more testing is required before using it in such larger domains. © Springer-Verlag Berlin Heidelberg 2013.