作者: Rui Wang , Günter Neumann
关键词:
摘要: We present a novel approach to RTE that exploits structure-oriented sentence representation followed by similarity function. The structural features are automatically acquired from tree skeletons extracted and generalized dependency trees. Our method makes use of limited size training data without any external knowledge bases (e.g. WordNet) or handcrafted inference rules. have achieved an accuracy 71.1% on the RTE-3 development set performing 10-fold cross validation 66.9% test data.