Textual entailment recognition based on dependency analisys and WordNet

Abstract

The UNED-NLP Group Recognizing Textual Entailment System is based on the use of a broad-coverage parser to extract dependency relations and a module which obtains lexical entailment relations from WordNet. The work aims at comparing whether the matching of dependency trees subestructures give better evidence of entailment than the matching of plain text alone.