International Workshop on Spoken Language Translation (IWSLT) 2008

Honolulu, Hawaii, USA
October 20-21, 2008

Improving Statistical Machine Translation by Paraphrasing the Training Data

Francis Bond (1), Eric Nichols (2), Darren Scott Appling (3), Michael Paul (1)

(1) National Institute of Information and Communications Technology, Japan
(2) Nara Institute of Science and Technology, Japan
(3) Georgia Institute of Technology, Atlanta, GA, USA

Large amounts of training data are essential for training statistical machine translations systems. In this paper we show how training data can be expanded by paraphrasing one side. The new data is made by parsing then generating using a precise HPSG based grammar, which gives sentences with the same meaning, but minor variations in lexical choice and word order. In experiments with Japanese and English, we showed consistent gains on the Tanaka Corpus with less consistent improvement on the IWSLT 2005 evaluation data.

Full Paper     Presentation (pdf)

Bibliographic reference.  Bond, Francis / Nichols, Eric / Appling, Darren Scott / Paul, Michael (2008): "Improving statistical machine translation by paraphrasing the training data", In IWSLT-2008, 150-157.