Extracting Hierarchical Rules from a Weighted Alignment Matrix
Zhaopeng Tu, Yang Liu, Qun Liu, Shouxun Lin · 2011
Word alignment is a fundamental step in machine translation. Current statistical machine translation systems suffer from a major drawback: they only extract rules from 1-best alignments, which adversely affects the rule sets quality due to alignment mistakes. To alleviate this problem, we extract hierarchical rules from weighted alignment matrix (Liu et al., 2009). Since the sub-phrase pairs would change the inside and outside areas in the weighted alignment matrix of the hierarchical rules, we propose a new algorithm to calculate the relative frequencies and lexical weights of hierarchical rules. To achieve a balance between rule table size and performance, we construct a scoring measure that incorporates both frequency and lexical weight to select the best target phrase for each source phrase. Experiments show that our approach improves BLEU score by ranging from 1.4 to 1.9 points over baseline for hierarchical phrase-based, and 1.4 to 1.5 points for tree-to-string model. 1