Causal Model Progressions as a Foundation for Intelligent Learning Environments

Abstract AI research in qualitative modeling makes possible new approaches to teaching people about science and technology. We are exploring the implications of this work for the design of intelligent learning environments. The domain of application is electrical circuits, but the approach can be generalized to other subjects. Our prototype instructional system is based upon a progression of qualitative models of electrical circuit behavior. These models enable the system to simulate circuit behavior and to generate causal explanations. They also serve as target mental models for the learner. The model progression is used to create problem sets that motivate successive refinements to the students' mental models. Acquisition of these models allows students, at all stages of learning, to solve interesting problems, such as circuit design and troubleshooting problems. The system enables students to employ different learning strategies and to manage their own learning. For instance, they can create and experiment with circuits, can attempt problems posed by the system, and can ask for feedback and coaching from the models. In pilot trials, the learning environment successfully taught novices to troubleshoot and to mentally simulate circuit behavior.