Journal of Engineering Education Transformations

Abstract

Abstract: Educational infrastructure is moving towards rapid digitization to conduct and evaluate examinations for remote students. Many universities now offer globally recognized distance learning courses to cater to a wider audience. However, this transition comes with its set of challenges, particularly for professors and staff members who find themselves burdened with a substantial amount of manual work during the examination season. The tasks include setting up unique question papers for every exam, including different types of questions with varying difficulty, and eventually evaluating the answers given by the students, which is not only timeconsuming but also a labour-intensive process. To address this issue, the paper proposes a solution that aims to reduce the workload of teaching staff by enhancing the efficiency of the examination process. It does so by leveraging several natural language processing techniques for generating two types of questions- objective and subjective, and grading the solutions of the examinee. Additionally, subjective questions are further classified based on Bloom's taxonomy levels, providing a diverse range of questions that align with varying cognitive abilities. The automation of this process not only eases the burden on educators but also ensures a more streamlined and effective examination process, thus contributing to the broader goal of digitizing education.

Keywords: Automated Question Generation, Bloom's Taxonomy, Workload Reduction.