Section 3 contains an explanation of our bilingual extension to the word2vec model builder.
It is followed by a section on experimental results for word-aligned English-Arabic and English-Hebrew corpora and then a brief discussion of the implications.