A hierarchical Dirichlet language model

作者: David J. C. MacKay , Linda C. Bauman Peto

DOI: 10.1017/S1351324900000218

关键词:

摘要: We discuss a hierarchical probabilistic model whose predictions are similar to those of the popular language modelling procedure known as 'smoothing'. A number interesting differences from smoothing emerge. The insights gained view this problem point towards new directions for modelling. ideas paper also applicable other problems such triphomes in speech, and DNA protein sequences molecular biology. algorithm is compared with on two million word corpus. methods prove be about equally accurate, using fewer computational resources.

参考文章(23)
Radford M. Neal, Bayesian Mixture Modeling Springer, Dordrecht. pp. 197- 211 ,(1992) , 10.1007/978-94-017-2219-3_14
John Skilling, Classic Maximum Entropy Springer Netherlands. pp. 45- 52 ,(1989) , 10.1007/978-94-015-7860-8_3
David J. C. MacKay, Hyperparameters: Optimize, or Integrate Out? Maximum Entropy and Bayesian Methods. pp. 43- 59 ,(1996) , 10.1007/978-94-015-8729-7_2
Stephen F. Gull, Developments in Maximum Entropy Data Analysis Springer, Dordrecht. pp. 53- 71 ,(1989) , 10.1007/978-94-015-7860-8_4
David Nahamoo, Lalit R. Bahl, Peter V. de Souza, Robert L. Mercer, Peter F. Brown, A fast algorithm for deleted interpolation. conference of the international speech communication association. ,(1991)
David S Touretzky, Jeffrey L Elman, Terrence J Sejnowski, Connectionist models : proceedings of the 1990 summer school M. Kaufmann Publishers. ,(1991)
William A. Gale, Kenneth W. Church, A program for aligning sentences in bilingual corpora Computational Linguistics. ,vol. 19, pp. 75- 102 ,(1993) , 10.5555/972450.972455
John Stutz, Peter Cheeseman, Robin Hanson, Bayesian classification with correlation and inheritance international joint conference on artificial intelligence. pp. 692- 698 ,(1991)
Linda C. Bauman Peto, A Comparison of Two Smoothing Methods for Word Bigram Models arXiv: Computation and Language. ,(1994)
Christopher K.I. Williams, Geoffrey E. Hinton, Mean field networks that learn to discriminate temporally distorted strings Connectionist Models#R##N#Proceedings of the 1990 Summer School. pp. 18- 22 ,(1991) , 10.1016/B978-1-4832-1448-1.50008-1