New Perspectives in Science Education

The integration of AI into the educational process is the subject of this study. Particular attention is paid to the technology of creating an up-to-date training sample in the process of use. For this purpose, a hybrid-intelligent bot (HI) under the guise of a student is anonymously integrated into study groups and included in the educational process. The initial training of the bot is carried out on the basis of the program of the current educational course. Then, based on the analysis of the results of regular (after each cycle of lectures) "cross-testing" of participants, additional training of the bot is carried out. The most important feature of cross-testing [1], which allows you to quantitatively and separately evaluate creative and analytical abilities, as well as the ability to make management decisions and to implement them, etc. The HI-bot makes tests on the current topic, which are solved by other students and, in turn, answers the tests made by them. Cross-tests are based on the methods of "fuzzy logic". Each test consists of a task that has an "indefinite" number of correct answer options. The score for the answer varies from 0 to 1, depending on how many options were indicated by the respondent. In addition, the number of test participants who solved the task set in the test and in what time is taken into account. Thus, the assessment of the results of each student consists of at least two indicators. One of them allows you to assess the ability to set tasks, the other - to implement them. The HI-bot is trained on correct and erroneous examples. The process of forming a training sample takes into account the various opinions of the participants. The HI-bot effectively and quickly develops during use, transferring the accumulated experience from one user to another. In the process of communication, thematic orientation of AI arises due to the formation of a new training sample, complementing the existing one and training on its basis. This ensures the specialization of the HI system, which integrates people and AI. Thus, the project implements the basic idea of integrating AI with human intelligence and creates the prerequisites for "training" and "individual education" of HI under the influence of human intelligence.

Keywords: Artificial Intelligence, Hybrid Intelligence, fuzzy logic, cross-testing

REFERENCE:

[1] S.Gorelik, V.Grudinin, E. Ishutina. International Research Journal, No 7 (145) July, https://r esearch-journal.org/media/artic les/13463_review.pdf