Abstract:

Artificial intelligence, machine learning and expert systems are IT tools that play a significant role in managing the effective acquisition of individual knowledge. Currently, a number of intelligent systems to support learning are operating in domestic and international practice. Most of them are course-oriented, that is, they create a field of knowledge by organizing the subject matter to be mastered into a logical structure. During ongoing cycles of learning and assessment, the student is individually guided and helped to build his own knowledge system in the given topic.

The article discusses the methodology of step-by-step solving of mathematical problems with ITS inner loop support using ChatGPT4+Wolfram Alpha. Using prepared examples, we present a methodology by which a teacher can use ChatGPT4 + Wolfram Alpha — with pre-designed prompts — to prepare inner loop of an intelligent tutoring system to handle a lesson or topic from secondary school mathematics. Step-by-step tutoring for solving tasks is jointly carried out by the language model and WA. The two connected intelligent systems can correct errors that normally occur during solving, as well as provide the student with suggestions on how to proceed and continue with partially solved tasks. In this way the solution of the task is successfully completed and thus the student acquires the necessary skills in an active and independent learning process. The presented methodology deals with the curriculum of secondary school mathematics.

Keywords:

ITS inner loop, stepwise exercise, feedback service based on ChatGPT4 and Wolfram Alpha, prompt engineering.