AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频


pub
生成解读视频

AI 溯源

AI解析本论文相关学术脉络


Master Reading Tree
生成 溯源树

AI 精读

AI抽取本论文的概要总结


微博一下
Statistical machine translation systems within this framework typically consist of three components: a language model that assigns a propability P(e) to any given English string e; a translation model that assigns a probability P(f | e) to any given pair of English and French str...

Fast and optimal decoding for machine translation

Artif. Intell., no. 1-2 (2004): 127-143

引用357|浏览36
EI

摘要

A good decoding algorithm is critical to the success of any statistical machine translation system. The decoder's job is to find the translation that is most likely according to a set of previously learned parameters (and a formula for combining them). Since the space of possible translations is extremely large, typical decoding algorithm...更多

代码

数据

简介
  • For arbitrary word-reordering, the decoding problem is NP-hard (Knight, 1999)
  • It is a sensible, albeit still computationally intensive strategy to restrict the search to a large subset of likely decodings and choose just among them (Brown et al, 1995; Wang and Waibel, 1997).
  • As Wang and Waibel (1997) remark, it is hard to determine search errors — the only way to show that a decoding is sub-optimal is to produce a higher-scoring one
重点内容
  • Statistical machine translation (SMT) in the tradition of Brown et al (e.g. 1990, 1993), which is often referred to as the noisy channel approach to machine translation, restates the problem of finding the optimal English translation eof a French sentence f, or e = arg maxe P(e | f), as finding e = arg max P(f | e) · P(e).1 e

    Statistical machine translation systems within this framework typically consist of three components: (1) a language model (LM) that assigns a propability P(e) to any given English string e; (2) a translation model (TM) that assigns a probability P(f | e) to any given pair of English and French strings e and f; and (3) a decoding algorithm to perform the search.

    If the source and target languages are constrained to have the same word order, a linear Viterbi algorithm can be applied (Tillmann et al, 1997)
  • This paper reports on measurements of speed, search errors, and translation quality in the context of a traditional stack decoder (Jelinek, 1969; Brown et al, 1995) and two new decoders
  • The experiments reported were set up so that all decoders worked on the same search space
  • For the experiments reported in Tab. 1, we used a bigram language model
  • The results reported in Tab. 2 were obtained using a trigram model
结论
  • The experiments reported were set up so that all decoders worked on the same search space.
  • The integer programming decoder explores this space exhaustively.
  • The stack and greedy decoders explore only a portion of it.
  • For the experiments reported in Tab. 1, the authors used a bigram language model.
  • The results reported in Tab. 2 were obtained using a trigram model
表格
  • Table1: Comparison of decoders on sets of 101 test sentences. All experiments in this table use a bigram language model. Translation errors can be syntactic, semantic, or both. Errors were counted on the sentence level, so that every sentence can have at most one error in each category
  • Table2: Comparison between decoders using a trigram language model. Greedy∗ and greedy1 are greedy decoders optimized for speed
Download tables as Excel
基金
  • This work was supported by DARPA-ITO grant N66001-00-1-9814
引用论文
  • Brown, P., Cocke, J., Della Pietra, S., Della Pietra, V., Jelinek, F., Lafferty, J., Mercer, R., Roossin, P., 1990. A statistical approach to machine translation. Computational Linguistics 16 (2), 79–85.
    Google ScholarLocate open access versionFindings
  • Brown, P., Cocke, J., Della Pietra, S., Della Pietra, V., Jelinek, F., Lai, J., Mercer, R., 1995. Method and system for natural language translation. US Patent 5,477,451.
    Google ScholarFindings
  • Brown, P., Della Pietra, S., Della Pietra, V., Mercer, R., 199The mathematics of statistical machine translation: Parameter estimation. Computational Linguistics 19 (2), 263–311.
    Google ScholarLocate open access versionFindings
  • Garey, M., Johnson, D., 1979. Computers and Intractability. A Guide to the Theory of NP-Completeness. W.H. Freeman and Company, New York.
    Google ScholarFindings
  • Hart, P. E., Nilsson, N. J., Raphael, B., 1968. A formal basis for the heuristic determination of minimum cost paths. IEEE Transactions on System Science and Cybernetics 4 (2), 100–107.
    Google ScholarLocate open access versionFindings
  • Jelinek, F., 1969. A fast sequential decoding algorithm using a stack. IBM Research Journal of Research and Development 13, 675–685.
    Google ScholarLocate open access versionFindings
  • Knight, K., 1999. Decoding complexity in word-replacement translation models. Computational Linguistics 25 (4), 607–615.
    Google ScholarLocate open access versionFindings
  • Monasson, R., Zecchina, R., Kirkpatrick, S., Selman, B., Troyansky, L., 1999. Determining computational complexity from characteristic ‘phrase transitions’. Nature 800 (8), 133–137.
    Google ScholarLocate open access versionFindings
  • Och, F., Ueffing, N., Ney, H., 2001. An efficient A∗ search algorithm for statistical machine translation. In: Proceedings of the ACL Workshop on Data-Driven Machine Translation. Toulouse, France, pp. 55–62.
    Google ScholarLocate open access versionFindings
  • Papineni, K., Roukos, S., Ward, T., Henderson, J., Reeder, F., 2002. Corpus-based comprehensive and diagnostic MT evaluation: Initial Arabic, Chinese, French, and Spanish results. In: Proceedings of the Human Language Technology Conference. San Diego, CA, pp. 124–127.
    Google ScholarLocate open access versionFindings
  • Selman, B., Levesque, H., Mitchell, D., 1992. A new method for solving hard satisfiability problems. In: Proceedings of the 10th National Conference on Artificial Intelligence. San Jose, California, pp. 440–446.
    Google ScholarLocate open access versionFindings
  • Tillmann, C., Vogel, S., Ney, H., Zubiaga, A., 1997. A DP-based search using monotone alignments in statistical translation. In: Proceedings of the 35th ACL. pp. 289–296.
    Google ScholarLocate open access versionFindings
  • Wang, Y., Waibel, A., 1997. Decoding algorithm in statistical machine translation. In: Proceedings of the 35th ACL. pp. 366–372.
    Google ScholarLocate open access versionFindings
  • Wu, D., 1996. A polynomial-time algorithm for statistical machine translation. In: Proceedings of the 34th ACL. pp. 152–158.
    Google ScholarLocate open access versionFindings
0
您的评分 :

暂无评分

标签
评论
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn