Dear students,
the exercise on token-level transfer was just made available. It is due by 30.06. 2pm.
We are aware that lecture and exercise are too closely aligning, which is why the exercise is designed such that you can solve it by having a decent grasp of the suggested readings.
We will discuss the generation and quality of commonly based datasets and how machine translation can help improve cross-lingual transfer in token-level tasks.
Best,
Fabian