Abstract
Maximum entropy (ME) models have been successfully applied to many natural language problems. In this paper, we show how to integrate ME models efficiently within a maximum likelihood training scheme of statistical machine translation models. Specifically, we define a set of context-dependent ME lexicon models and we present how to perform an efficient training of these ME models within the conventional expectation-maximization (EM) training of statistical translation models. Experimental results are also given in order to demonstrate how these ME models improve the results obtained with the traditional translation models. The results are presented by means of alignment quality comparing the resulting alignments with manually annotated reference alignments.
Keywords
- Target Word
- Training Corpus
- Translation Model
- Statistical Machine Translation
- Word Class
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
This work has been partially supported by Spanish CICYT under grant TIC2000-1599-C02-01
This is a preview of subscription content, access via your institution.
Buying options
Preview
Unable to display preview. Download preview PDF.
References
Berger, A.L., Della Pietra, S.A., Della Pietra, V.J.: A maximum entropy approach to natural language processing. Computational Linguistics 22 (1996) 39–72
Brown, P.F., Della Pietra, S.A., Della Pietra, V.J., Mercer, R.L.: The mathematics of statistical machine translation: Parameter estimation. Computational Linguistics 19 (1993) 263–311
Darroch, J., Ratcliff, D.: Generalized iterative scaling for log-linear models. Annals of Mathematical Statistics 43 (1972) 95–144
Della Pietra, S.A., Della Pietra, V.J., Lafferty, J.: Inducing features in random fields. IEEE Trans. on PAMI 19 (1997) 380–393
Foster, G.: Incorporating position information into a maximum entropy/minimum divergence translation model. In: Proc. of CoNNL-2000 and LLL-2000, Lisbon, Portugal (2000) 37–52
García-Varea, I., Och, F.J., Ney, H., Casacuberta, F.: Refined lexicon models for statistical machine translation usign a maximum entropy approach. In: Proc. of the 39th Annual Meeting of the ACL, Toulouse, France (2001) 204–211
Och, F.J.: An efficient method for determining bilingual word classes. In: 9th Conf. of the Europ. Chapter of the ACL, Bergen, Norway (1999) 71–76
Och, F.J., Ney, H.: Giza++: Training of statistical translation models (2001) http://www-i6.Informatik.RWTH-Aachen.DE/~och/software/GIZA++.html.
Och, F.J., Ney, H.: A comparison of alignment models for statistical machine translation. In: COLING’ 00: The 18th Int. Conf. on Computational Linguistics, Saarbrücken, Germany (2000) 1086–1090
Papineni, K., Roukos, S., Ward, R.: Maximum likelihood and discriminative training of direct translation models. In: Proc. Int. Conf. on Acoustics, Speech, and Signal Processing. (1998) 189–192
Rosenfeld, R.: A maximum entropy approach to adaptive statistical language modeling. Computer, Speech and Language 10 (1996) 187–228
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2002 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Varea, I.G., Och, F.J., Ney, H., Casacuberta, F. (2002). Efficient Integration of Maximum Entropy Lexicon Models within the Training of Statistical Alignment Models. In: Richardson, S.D. (eds) Machine Translation: From Research to Real Users. AMTA 2002. Lecture Notes in Computer Science(), vol 2499. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45820-4_6
Download citation
DOI: https://doi.org/10.1007/3-540-45820-4_6
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-44282-0
Online ISBN: 978-3-540-45820-3
eBook Packages: Springer Book Archive