A Novel Reordering Model Based on Multi-layer Phrase for Statistical Machine Translation

Phrase reordering is of great importance for statistical machine translation. According to the movement of phrase translation, the pattern of phrase reordering can be divided into three classes: monotone, BTG (Bracket Transduction Grammar) and hierarchy. It is a good way to use different styles of reordering models to reorder different phrases according to the characteristics of both the reordering models and phrases itself. In this paper a novel reordering model based on multi-layer phrase (PRML) is proposed, where the source sentence is segmented into different layers of phrases on which different reordering models are applied to get the final translation. This model has some advantages: different styles of phrase reordering models are easily incorporated together; when a complicated reordering model is employed, it can be limited in a smaller scope and replaced with an easier reordering model in larger scope. So this model better trade-offs the translation speed and performance simultaneously.