Research on Deep Learning HMM Word Alignment

被引:0
作者
Li, Dan [1 ]
Yu, Zheng-hong [2 ]
机构
[1] Coll Post & Telecommun WIT, Wuhan, Peoples R China
[2] Wuhan Univ Sci & Technol, City Coll, Wuhan, Peoples R China
来源
INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE: TECHNIQUES AND APPLICATIONS, AITA 2016 | 2016年
关键词
Deep learning; Word Alignment; HMM;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Word alignment is an essential step for machine translation. According to the over-fitting of traditional word alignment model and the weakness for the context description, the deep learning hidden HMM word alignment combined a multi-layer neural network with an undirected probabilistic graph model, use the similarity of the word and context information word alignment to be a more precise modeling. Experimental results show that, compared with the reference system, this model can significantly improve the effect of word alignment, which is applicable.
引用
收藏
页码:139 / 143
页数:5
相关论文
共 13 条
  • [1] [Anonymous], 1996, COMPUTATIONAL LINGUI
  • [2] [Anonymous], 2010, Statistical Machine Translation
  • [3] Auli Michael, 2013, P 2013 C EMPIRICAL M, P1044
  • [4] Deng Y, 2008, AUDIO SPEECH LANGUAG
  • [5] Devlin J, 2014, P ACL 2014 IN PRESS
  • [6] Feng Y, 2013, P P ACL
  • [7] Haghighi A, 2009, P P ACL IJCNLP
  • [8] He X., 2007, P 2 WORKSH STAT MACH
  • [9] Deep Neural Networks for Acoustic Modeling in Speech Recognition
    Hinton, Geoffrey
    Deng, Li
    Yu, Dong
    Dahl, George E.
    Mohamed, Abdel-rahman
    Jaitly, Navdeep
    Senior, Andrew
    Vanhoucke, Vincent
    Patrick Nguyen
    Sainath, Tara N.
    Kingsbury, Brian
    [J]. IEEE SIGNAL PROCESSING MAGAZINE, 2012, 29 (06) : 82 - 97
  • [10] Liang Percy, 2006, P NAACL