Machine Translation (MTAT.06.055)
The course is meant for anyone who wants to
- find out how machine translation works, why it makes funny mistakes
- which algorithms and resources it requires
- what its strong and weak sides are
- where and how it is applied in practice
Most importantly, as part of the course you will build state-of-the-art statistical and neural machine translation systems. As prerequisite we expect you to be very comfortable with algorithms, programming and to not be afraid to learn some more math, applied probability theory, machine learning and a tiny bit of linguistics.
There will be lectures done by Mark almost every week on Wednesday at 10:15 in Liivi 2 - 612. As the practical side of the course you will do a shared task project, as part of which everyone will build a translation system using the same train/dev/test dataset for the same language pair, using your own devised methods of improving the translation quality (and winning the shared task? >:-) Shared task supervision and some introductory labs are given by Maksym on some Mondays at 12:15 in Liivi 2 - 612.