I Do Not Understand What I Cannot Define: Automatic Question Generation With Pedagogically-Driven Content Selection

Tim Steuer,Anna Filighera,Tobias Meuser,Christoph Rensing
DOI: https://doi.org/10.48550/arXiv.2110.04123
2021-10-08
Abstract:Most learners fail to develop deep text comprehension when reading textbooks passively. Posing questions about what learners have read is a well-established way of fostering their text comprehension. However, many textbooks lack self-assessment questions because authoring them is timeconsuming and expensive. Automatic question generators may alleviate this scarcity by generating sound pedagogical questions. However, generating questions automatically poses linguistic and pedagogical challenges. What should we ask? And, how do we phrase the question automatically? We address those challenges with an automatic question generator grounded in learning theory. The paper introduces a novel pedagogically meaningful content selection mechanism to find question-worthy sentences and answers in arbitrary textbook contents. We conducted an empirical evaluation study with educational experts, annotating 150 generated questions in six different domains. Results indicate a high linguistic quality of the generated questions. Furthermore, the evaluation results imply that the majority of the generated questions inquire central information related to the given text and may foster text comprehension in specific learning scenarios.
Computation and Language
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to automatically generate educationally meaningful questions to enhance students' understanding of higher - education textbooks. Specifically, the authors are concerned that most students are unable to develop a deep understanding ability when passively reading textbooks, and asking questions is an effective method to promote text understanding. However, many textbooks lack self - assessment questions because writing these questions is both time - consuming and expensive. An automatic question generator can alleviate this problem, but generating high - quality educational questions faces linguistic and pedagogical challenges. The paper proposes an automatic question generator based on learning theory, aiming to generate meaningful questions by selecting content with educational value, thereby promoting students' text understanding and learning effectiveness. The research question (RQ) of the paper is: "To what extent can educationally meaningful questions be automatically generated for higher - education textbooks?" The authors verified the quality of the questions generated by their method in textbooks of different disciplines through experiments, including language quality, relevance to the central information of the text, and educational value.