When you did the supervised alignment, did you try to align all languages together, or did you align them by pair (english/french, english/Spanish, english/italian ...) and maybe taking the english vector space as some kind of universal one.
The reason I'm asking is because I downloaded the multilingual embeddings and for some pair of languages, there are words you would really expect to see close to each other but they are actually quite far from one another.
This is not much of an issue but more a question.
When you did the supervised alignment, did you try to align all languages together, or did you align them by pair (english/french, english/Spanish, english/italian ...) and maybe taking the english vector space as some kind of universal one.
The reason I'm asking is because I downloaded the multilingual embeddings and for some pair of languages, there are words you would really expect to see close to each other but they are actually quite far from one another.
Thanks!