Embedding process knowledge in cyber-social systems on the example of cognitive tutor to teach scope

Abstract

Goal. Studying the effects of embedding human knowledge in intelligent tutoring systems in the context of cyber social systems. Problem. The usage of human- and machine-readable subject-domain models such as thought process graphs for building intelligent tutoring systems for the cognitive knowledge domain on the example of a tutor to teach variable scope. Methodology. In this study, we consider the structure and interactions of system components of the CompPrehension intelligent tutoring system. Then we explain the method of embedding human knowledge into it by constructing a subject domain model and demonstrate it on the example of the model for determining variable scopes. Then we provide results of the preliminary evaluation of the created tutor. Results. The learning gains of study participants after using the tutor for about 12 minutes were statistically significant, which proves that the tutor is effective. The learners who often repeated their errors in subsequent learning problems had smaller learning gains than the other learners; the average time-per-learning problem did not affect the learning gains. The results of the usability survey were positive. Implications. Using cognitive modeling by developing thought process graphs for building intelligent tutoring systems opens the way not just to building effective tutors, but also allows to solve advanced tasks like generating learning problems and pedagogical dialogue, which significantly expand the features of intelligent tutoring systems and bring them closer to the functions of human tutors.