Using AI Cognitive Diagnosis Embedded in a Game-Based Learning Environment

The aim of this study was to develop a game-based environment with embedded AI cognitive diagnosis to deliver personalized instruction to learners. The system identified participants' possible misconceptions using a Bayesian network. Corresponding learning games based on their misconceptions were then provided to the learners. This research compared the effects of AI cognitive diagnosis in a game-based learning environment to those in a multimedia learning environment. A quasi-experimental design was used in the experiments, with fifty-six secondary school students as participants. The results showed that the accuracy of the concept diagnosis system was 90.8% (expert judgment). The personalized learning outcomes of students in the game-based learning environment were superior to those engaged in multimedia problem-solving tasks. The results implied that the game-based learning with cognitive diagnosis mechanism plays an important role in enhancing problem-solving skills, learning motivation, misconception correction, and flow experience.