Fast BTG-Forest-Based Hierarchical Sub-sentential Alignment

In this paper, we propose a novel BTGforest-based alignment method. Based on a fast unsupervised initialization of parameters using variational IBM models, we synchronously parse parallel sentences top-down and align hierarchically under the constraint of BTG. Our twostep method can achieve the same run-time and comparable translation performance as fast align while it yields smaller phrase tables. Final SMT results show that our method even outperforms in the experiment of distantly related languages, e.g., English–Japanese.