中国科学院自动化研究所机构知识库

This paper proposes a well-formed dependencyto string translation model with BTG grammar. By enabling the usage of wellformed sub-structures and allowing flexible reordering of them, our approach is effective to relieve the problems of parsing error and flatness in dependency structure. To utilize the well-formed dependency rules during decoding, we adapt the tree traversal decoding algorithm into a bottom-up CKY algorithm. And a lexicalized reordering model is used to encourage the proper combination of two neighbouring blocks. Experiment results demonstrate that our approach can effectively improve the performance by more than 2 BLEU score over the baseline.