Model-Based Search for Statistical Machine Translation
Jiefan Zhang · 2005
I address the complexity problem of phrase-base translation model. I view the problem of machine translation as to find compact models having Minimum Description Length (MDL). Then I apply model merging on translation model to find compact models. This method uses the Joint-Probability Phrase-base Model as the initial model, and then gradually eliminates redundant translation entries according to the MDL princi-ple. The size of result translation model is only 1 % of the initial model, along with comparable performance. i Acknowledgements First, I would like to thank my supervisor Miles Osborne for his help, supervision and understanding throughout the project in this year. I would also like to thank Philipp Koehn and Markus Becker for their help. Finally, I would like to thank my family and friends for being supportive while I study overseas. ii