Training Data in Statistical Machine Translation - the More, the Better?

Current statistical machine translation (SMT) systems are stated to be dependent on the availability of a very large training data for producing the language and translation models. Unfortunately, large parallel corpora are available for a limited set of language pairs and for an even more limited set of domains. In this paper we investigate the behavior of an SMT system exposed to training data of different sizes and types. Our experimental results show that even parallel corpora of modest sizes can be used for training purposes without lowering too much the evaluation scores. We consider two language pairs in both translation directions for the experiments: English-Romanian and German-Romanian.