摘要
扩展HMM模型可以解决词语对齐结果与句法约束冲突,从而更好地进行词语对齐。在短语对齐基础上利用目标语言的短语结构树抽取翻译规则。采用扩展CYK算法CYKA+作为系统的解码器,该算法可以处理非乔姆斯基范式的翻译规则;采用两轮解码算法在解码过程中整合语言模型。实验表明,与传统词语对齐模型相比,改进的HMM词语对齐模型具有更高的对齐准确率,并且翻译结果的BLEU评测得分更高。采用翻译规则的系统在不同数据集上具有更稳定的翻译结果。两轮解码算法与立方剪枝算法具有相近的解码质量,但前者解码速度更快。
Improved hidden Markov model was used to align words and solve the inconsistency between word alignment and phrase structures. Translation rules were extracted based on aligned phrases and English phrase trees. An extended CYK CYK＊ algorittun was used as the decoder and a two-pass-decoding algorithm was proposed for intergrating the language model during decoding, which can decode non Chomsky normal form. The experimental results show the BLEU score of improved HMM is higher than the score of HMM, and the translation quality of translation rules is bet ter than phrase-based machine transtion. The BLEU score of two-pass-decoding algorithm is close to the score ot cube prune algorithm and decoding time costs less.
出处
《计算机科学》
CSCD
北大核心
2013年第2期214-217,共4页
Computer Science
基金
教育部自主科研项目(20111081010)资助
