Syntactic Constraints on Paraphrases Extracted from Parallel Corpora

We improve the quality of paraphrases extracted from parallel corpora by requiring that phrases and their paraphrases be the same syntactic type. This is achieved by parsing the English side of a parallel corpus and altering the phrase extraction algorithm to extract phrase labels alongside bilingual phrase pairs. In order to retain broad coverage of non-constituent phrases, complex syntactic labels are introduced. A manual evaluation indicates a 19% absolute improvement in paraphrase quality over the baseline method.

• Type: Conference
• Year: 2008
• Where: EMNLP
• Authors: Chris Callison-Burch
• Keywords: Bilingual Phrase Pairs, EMNLP 2008, Natural Language Processing, Paraphrase, Phrase Extraction Algorithm