Generating Questions for German Text

Not very often, but every now and then, I receive a request for my master's thesis from 2015, so I thought it might be useful to make it publicly available here. It has the same title as this post and describes the implementation and evaluation of a rule-based question generation system for German text.

I was curious why people still care about this system and found a few recent publications from the CL group around my old professor in Tübingen that refer to it.

• De Kuthy, K., Kannan, M., Ponnusamy, H. S., & Meurers, D. (2020). Towards automatically generating questions under discussion to link information and discourse structure. In Proceedings of the 28th International Conference on Computational Linguistics (pp. 5786-5798).
• Kannan, M., Ponnusamy, H. S., De Kuthy, K., Stein, L., & Meurers, D. (2021). Exploring Input Representation Granularity for Generating Questions Satisfying Question-Answer Congruence. In Proceedings of the 14th International Conference on Natural Language Generation (pp. 24-34).
• Ponnusamy, H. S., Kannan, M., De Kuthy, K., & Meurers, D. (2021). Advancing Neural Question Generation for Formal Pragmatics: Learning when to generate and when to copy. In Integrating Perspectives on Discourse Annotation (DiscAnn).
• De Kuthy, K., Kannan, M., Santhi Ponnusamy, H., & Meurers, D. (2022). Exploring Neural Question Generation for Formal Pragmatics: A New Data Set and Comparison of Models. Frontiers in Artificial Intelligence, 233.

Some chapters of the thesis are still interesting to read. The weakest parts are the evaluation, which I deferred as much as possible to have more time to work on the system, and the software engineering standards of the implementation, which is why I can't offer a working implementation anymore and it took quite some back and forth at the beginning of 2018 to make it work for Kordula de Kuthy in Tübingen. So if you're interested in running the system, you'll have to ask the group in Tübingen whether they'd share their updated version with you.