作者: Günter Neumann , Rui Wang
DOI:
关键词:
摘要: We present a novel approach to recognizing Textual Entailment. Structural features are constructed from abstract tree descriptions, which automatically extracted syntactic dependency trees. These then applied in subsequence-kernel-based classifier learn whether an entailment relation holds between two texts. Our method makes use of machine learning techniques using limited data set, no external knowledge bases (e.g. WordNet), and handcrafted inference rules. achieve accuracy 74.5% for text pairs the Information Extraction Question Answering task, 63.6% RTE-2 test data, 66.9% RET-3 data.