A Feasibility Study of Answer-Agnostic Question Generation for Education
This addresses the issue of generating relevant questions from textbook passages for educational purposes, but it is incremental as it focuses on feasibility and error reduction.
The study tackled the problem of answer-agnostic question generation for education by showing that using summarized input, especially human-written summaries, significantly improves question acceptability from 33% to 83%.
We conduct a feasibility study into the applicability of answer-agnostic question generation models to textbook passages. We show that a significant portion of errors in such systems arise from asking irrelevant or uninterpretable questions and that such errors can be ameliorated by providing summarized input. We find that giving these models human-written summaries instead of the original text results in a significant increase in acceptability of generated questions (33% $\rightarrow$ 83%) as determined by expert annotators. We also find that, in the absence of human-written summaries, automatic summarization can serve as a good middle ground.