Question Taxonomy and Implications for Automatic Question Generation

Authors: 

Ming Liu and Rafael A. Calvo

Paper type: 
Poster

Abstract: 

Many Automatic Question Generation (AQG) approaches have been proposed focusing on reading comprehension support; however, none of them addressed academic writing. We conducted a large-scale case study with 25 supervisors and 36 research students enroled in an Engineering Research Method course. We investigated trigger questions, as a form of feedback, produced by supervisors, and how they support these students’ literature review writing. In this paper, we identified the most frequent question types according to Graesser and Person’s Question Taxonomy and discussed how the human experts generate such questions from the source text. Finally, we proposed a more practical Automatic Question Generation Framework for supporting academic writing in engineering education.

Keywords: 

Academic Writing Support, Question Taxonomy, Question Generation