BTG-based machine translation with simple reordering model using structured perceptron

In this paper, we present a novel statistical machine translation method which employs a BTG-based reordering model during decoding. BTG-based reordering models for preordering have been widely explored, aiming to improve the standard phrase-based statistical machine translation system. Less attention has been paid to incorporating such a reordering model into decoding directly. Our reordering model differs from previous models built using a syntactic parser or directly from annotated treebanks. Here, we train without using any syntactic information. The experiment results on an English-Japanese translation task show that our BTG-based decoder achieves comparable or better performance than the more complex state-of-the-art SMT decoders.