``English''-English phrase table

``English''-English phrase tables are constructed as follows.

1. Parallel Corpus

   We prepare a Japanese-English parallel corpus.

2. Pattern-Based Machine Translation

   We use Japanese-English PBMT. Thus, we obtain ``English'' sentences from Japanese sentences. These ``English'' sentences are pairs of English sentences.

3. ``English''-English phrase tables

   We construct ``English''-English phrase tables using Giza++ [Och and Ney2003] and train-model.perl [Koehn
   
   2007] from the ``English'' sentences (outputs of Japanese-English PBMT) and English sentences (from the parallel corpus).

Figure 3 is a flow chart that shows how ``English''-English phrase tables are constructed.

Figure: Flowchart for constructing ``English''-English Phrase Tables