Adapting a Hybrid Intelligent Reinforcement Learning Environment

• Main Authors: Pavel Basalin, Dmitrii Kulikov, Yuliya Maskina
• Format: Article
• Language: Russian
• Published: The Fund for Promotion of Internet media, IT education, human development «League Internet Media» 2020-11-01
• Series: Современные информационные технологии и IT-образование
• Subjects: hybrid intelligent learning environment; knowledge-based system; fuzzy production rules; feedforward neural network; reinforcement adaptation
• Online Access: http://sitito.cs.msu.ru/index.php/SITITO/article/view/692

Learning is seen as the process of interaction between two sides. First one is a virtual teacher - an intellectual learning environment that accumulates in its base the didactically and methodically structured material of the specific discipline for transferring it to the student. The other side is the student who is “absorbed” by his consciousness in the learning environment and actively perceived material transferred to him, i.e. not just putting it in his mind (memory), but also conducive to the rational organization of the learning process. This article outlines the principles of organizing a hybrid intelligent learning environment, integrating models based on knowledge of the production type, and neural network technologies for decision-making based on two learning strategies: the self-navigation strategy of the student through the discipline material and the strategy of his complete submission to the intellectual learning environment. This helps to support the important role of the student in the formation of learning scenario and facilitates the solution of the problems of adapting the intellectual learning environment to the individual characteristics of the student, not only adapting with reinforcement, but also with the teacher, to a certain extent, in the role of which the student acts.