Hacker News Reader

If the European country she is from is within the EU, you can probably get parallel corpus from the EU parliamentary translations. I believe there are other EU research projects which release useful corpus.

View on HN

Sadly one of the largest tranlation corpus of multiple languages can't be used for copyright reasons i.e the user made subtitle files that are shared all over the internet. Every week a few hundred episodes of different shows are released in different countries that are translated by users into many languages. Now that I think about it an app that could transcribe the words spoken in a TV show and then would translate it using google translate would be a huge with the manga crowd. Though Japanese hasn't been added yet.

It's actually not at all clear if they couldn't be used. Things like Google's word2vec model are derived from copyrighted data and seem to be fine.

Perhaps distribution of the subtitles is illegal under copyright law, but I suspect the trained model would be fine.

It would be an interesting case anyway.

This is a directory of available multilingual corpora:

http://opus.lingfil.uu.se/

Reply on news.ycombinator.com