Improving Lexical Semantics for Sentential Semantics: Modeling Selectional Preference and Similar Words in a Latent Variable Model

Sentence Similarity [SS] computes a similarity score between two sentences. The SS task differs from document level semantics tasks in that it features the sparsity of words in a data unit, i.e. a sentence. Accordingly it is crucial to robustly model each word in a sentence to capture the complete semantic picture of the sentence. In this paper, we hypothesize that by better modeling lexical semantics we can obtain better sentential semantics. We incorporate both corpus-based (selectional preference information) and knowledge-based (similar words extracted in a dictionary) lexical semantics into a latent variable model. The experiments show state-of-the-art performance among unsupervised systems on two SS datasets.