Abstract:

Text entailment is foundational for text understanding and semantic inference,and recently it has attracted considerable attention in the computational linguistics community.Text contradiction is a negation of text entailment and both of them are the key aspects of semantics.In this paper,we aim to address the problem of textual contradiction detection on the basis of textual entailment,and a textual contradiction recognition approach based on semantic rules is proposed to tackle this problem.Experimental results demonstrate the effectiveness and feasibility of the textual contradiction recognition approach,which can also significantly improve the performance of textual entailment recognition.

Key words: text entailment;contradiction;statistical feature;semantic rules