A Logic Prover Approach to Predicting Textual Similarity

This paper presents a logic prover approach to predict-ing textual similarity. Sentences are represented using three logic forms capturing different levels of knowl-edge, from only content words to semantic representa-tions extracted with an existing semantic parser. A logic prover is used to find proofs and derive semantic fea-tures that are combined in a machine learning frame-work. Experimental results show that incorporating the semantic structure of sentences yields better results than simpler pairwise word similarity measures.