Head-modifier Relation based Non-lexical Reordering Model for Phrase-Based Translation

Phrase-based statistical MT (SMT) is a milestone in MT. However, the translation model in the phrase based SMT is structure free which greatly limits its reordering capacity. To address this issue, we propose a non-lexical headmodifier based reordering model on word level by utilizing constituent based parse tree in source side. Our experimental results on the NIST ChineseEnglish benchmarking data show that, with a very small size model, our method significantly outperforms the base

• Type: Journal
• Year: 2010
• Where: COLING
• Authors: Shui Liu, Sheng Li, Tiejun Zhao, Min Zhang, Pengyuan Liu
• Keywords: COLING 2010, Computational Linguistics, Constituent Based Parse, Small Size Model, Translation Model