Automatic Question Generation in Multimedia-Based Learning

We investigate whether questions generated automatically by two Natural Language Processing (NLP) based systems (one developed by the authors, the other a state-of-the-art system) can successfully be used to assist multimedia-based learning. We examine the feasibility of using a Question Generation (QG) system’s output as pre-questions; with different types of pre-questions used: text-based and with images. We also compare the psychometric parameters of the automatically generated questions by the two systems and of those generated manually. Specifically, we analyse the effect such pre-questions have on test-takers’ performance on a comprehension test about a scientific video documentary. We also compare the discrimination power of the questions generated automatically against that of questions generated manually. The results indicate that the presence of pre-questions (preferably with images) improves the performance of test-takers. They indicate that the psychometric parameters of the questions generated by our system are comparable if not better than those of the state-of-the-art system.