A syntactic textual entailment system based on dependency parser

Abstract

The development of a syntactic textual entailment system that compares the dependency relations in both the text and the hypothesis has been reported. The Stanford Dependency Parser has been run on the 2-way RTE-3 development set and the dependency relations obtained for a text and hypothesis pair has been compared. Some of the important comparisons are: subject-subject comparison, subject-verb comparison, object-verb comparison and cross subject-verb comparison. Corresponding verbs are further compared using the WordNet. Each of the matches is assigned some weight learnt from the development corpus. A threshold has been set on the fraction of matching hypothesis relations based on the development set. The threshold score has been applied on the RTE-4 gold standard test set using the same methods of dependency parsing followed by comparisons. Evaluation scores obtained on the test set show 54.75% precision and 53% recall for YES decisions and 54.45% precision and 56.2% recall for NO decisions. © Springer-Verlag 2010.