×

A systematic comparison of various statistical alignment models. (English) Zbl 1234.68428

Summary: We present and compare various methods for computing word alignments using statistical or heuristic models. We consider the five alignment models presented in [P. F. Brown, S. A. Della Pietra, V. J. Della Pietra and R. L. Mercer, “The mathematics of statistical machine translation: parameter estimation”, ibid. 19, No. 2, 263–311 (1993)], the hidden Markov alignment model, smoothing techniques, and refinements. These statistical models are compared with two heuristic models based on the Dice coefficient. We present different methods for combining word alignments to perform a symmetrization of directed statistical alignment models. As evaluation criterion, we use the quality of the resulting Viterbi alignment compared to a manually produced reference alignment. We evaluate the models on the German-English Verbmobil task and the French-English Hansards task. We perform a detailed analysis of various design decisions of our statistical alignment system and evaluate these on training corpora of various sizes. An important result is that refined alignment models with a first-order dependence and a fertility model yield significantly better results than simple heuristic models. In the appendix, we present an efficient training algorithm for the alignment models presented.

MSC:

68T50 Natural language processing
62F10 Point estimation
PDF BibTeX XML Cite
Full Text: DOI

References:

[1] Baum L. E., Inequalities 3 pp 1– (1972)
[2] Brown Peter F, Computational Linguistics 19 (2) pp 263– (1993)
[3] Dempster A. P., N. M., Journal of the Royal Statistical Society, Series B 39 (1) pp 1– (1977)
[4] DOI: 10.2307/1932409
[5] Ker Sue J, Computational Linguistics 23 (2) pp 313– (1997)
[6] Knight Kevin, Computational Linguistics 25 (4) pp 607– (1999)
[7] DOI: 10.1162/089120100561683 · Zbl 01938517
[8] DOI: 10.1109/89.817451
[9] Smadja Frank, Computational Linguistics 22 (1) pp 1– (1996)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.