A Simple,Fast Strategy for Weighted Alignment Hypergraph
Weighted alignment hypergraph 4 is potentially useful for statistical machine translation, because it is the first study to simultane ously exploit the compact representation and fertility model of word alignment.Since estimating the probabilities of rules extracted from hypergraphs is an NP-complete problem, they propose a divide-and conquer strategy by decomposing a hypergraph into a set of independent subhypergraphs.However, they employ a Bulls algorithm to enumerate all consistent alignments for each rule in each subhypergraph, which is very time-consuming especially for the rules that contain non-terminals.This limits the applicability of this method to the syntax translation models, the rules of which contain many non-terminals (e.g.SCFG rules).In response to this problem, we propose an inside-outside algorithm to ef ficiently enumerate the consistent alignments.Experimental results show that our method is twice as fast as the Bulls algorithm.In addition, the efficient dynamic programming algorithm makes our approach applicable to syntax-based translation models.
statistical machine translation weighted alignment hyper graph optimization
Zhaopeng Tu Jun Xie Yajuan Lv Qun Liu
Department of Computer Science, University of California, Davis, USA ; Key Laboratory of Intelligent Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, CAS, Beijin Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, CAS, Beijin
国际会议
Second CCF Conference,NLPCC2013(第二届自然语言处理与中文计算会议)
重庆
英文
188-199
2013-11-15（万方平台首次上网日期，不代表论文的发表时间）