摘要
机器翻译译文质量的自动评价是推动机器翻译技术快速发展的一条重要途径。该文提出了基于List-MLE排序学习方法的译文自动评价方法。在此基础上,探讨引入刻画译文流利度和忠实度的特征,来进一步提高译文自动评价结果和人工评价结果的一致性。实验结果表明,在评价WMT11德英任务和IWSLT08BTEC CEASR任务上的多个翻译系统的输出译文质量时,该文提出的方法预测准确率高于BLEU尺度和基于RankSVM的译文评价方法。
Automatic evaluation of machine translation plays an important role in promoting the rapid development of machine translation. In this paper, we apply the ListMLE approach to learning to rank for machine translation auto- matic evaluation. In addition, we introduce the features of translation fluency and adequacy to further improve the consistency between the results of the automatic evaluation and human judgments. When assess the translation qual- ity of the submitted system outputs of WMT'll German-English tasks and IWSLT'08 BTEC CE ASR tasks, the ex- perimental results indicate that the predicted accuracy of the proposed approach is higher than the BLEU metric and the one based on RankSVM.
出处
《中文信息学报》
CSCD
北大核心
2013年第4期22-29,共8页
Journal of Chinese Information Processing
基金
国家自然科学基金资助项目(61203313
61272212
61163006)
江西省教育厅自然科学基金资助项目(GJJ12212)
关键词
机器译文评价
排序学习
ListMLE方法
人工评价
自动评价
machine translation evaluation
learning to rank
ListMLE approach
automatic evaluation
human eval-uation