Deep-Learning-Based Agile Teaching Framework of Software Development Courses in Computer Science Education

Compared with other courses, especially non-engineering ones, software development courses teaching (SDCT) has many unique characteristics, such as higher requirement to students’ logical reasoning, the need for students to have comprehensive problem-solving capabilities, good communication and teamwork skills, etc. The practices of Agile Methodology can be compatible with these characteristics. Therefore, it is a matter of course to integrate Agile Methodology with SDCT. This paper analyses the problems and difficulties in the practices of applying agile teaching in SDCT. In addition, it uses an agile teaching project graph instead of linear teaching schedule to control teaching progress. Moreover, a deep generative model is used to obtain the abstract representation of students’ implicit skill status so that the construction and implementation of agile teaching project graph can be carried out. Finally, a deep-learning-based agile teaching framework for SDCT is proposed, which has the potential to positively transform SDCT’s future.