PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

A phrase-based hidden semi-markov approach to machine translation
Jesús Andrés-Ferrer and Alfons Juan
In: 13th Annual Meeting of the European Association for Machine Translation, 14-15 May 2009, Barcelona, Spain.

Abstract

Statistically estimated phrase-based models promised to further the state-of-the-art, however, several works have shown that they behave worse than heuristically estimated phrase-based models. In this work we present a latent variable phrase-based translation model inspired on the hidden semi-Markov models, that does not degrade the system performance. Results show that this model incurs in an improvement over the baseline. Additionally, we show that both Baum-Welch and Viterbi training obtain the very same result, suggesting that the model gathers most of the probability mass into one bilingual segmentation.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
Other (BibTex File)
EPrint Type:Conference or Workshop Item (Poster)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Natural Language Processing
ID Code:5579
Deposited By:Alfons Juan
Deposited On:08 March 2010