Semi-automatic generation of textual exercises for software engineering education
Florian Huber, Georg Hagel · 2022 IEEE Global Engineering Education Conference (EDUCON) · 2022
Creating class diagrams from given texts is an important skill for software engineering students in higher computer science education. Choosing the right diagram components and arranging them correctly regularly challenges students. Therefore, a thoughtful teaching approach is required from educators. Creating such exercises for different courses with different contexts is time-consuming for educators. In preparation for an upcoming exam, students are limited to few available exercise texts. To address these challenges, a deep learning model was trained to automatically generate textual exercises for software engineering education. To verify whether the model is suitable for educational purposes or not, a study with software engineering students has been conducted. The results show, that the texts are well understood according to grammar and sentence structure. Also, students found the created exercise useful and would like to use comparable exercises for exam preparation.