XVIII Международная студенческая научная конференция Студенческий научный форум - 2026

1. Review of research and problem statement

In recent years, the use of artificial intelligence in the educational environment has been actively studied. The works of Lande D.V. consider the possibilities of AI for automating learning and supporting students. The fundamental works of Goodfellow, Bengio, and Courville describe the architectures of neural networks, but the practical use of chatbots to solve problems of higher mathematics is not sufficiently covered. Most of the research is devoted to classical computer mathematics systems (MATLAB, Maple), while dialogue neural network models are poorly studied.

This article is aimed at filling this gap and occupies an applied place among existing research, offering a comparative analysis of two AI systems in solving real educational problems.

2. Analysis of problem solutions

Task No 1. Determinants of the third order.

Both services correctly calculated the 3rd order determinant and presented a step-by-step solution. The answers received coincided with the reference values.

Fig. 1. Calculation of determinant in DeepSeek

Micro-inference: AI efficiently handles basic computational tasks in linear algebra.

Task No 2. Vector algebra.

A geometric interpretation was required for the solution. ChatGPT provided a detailed solution and a graphical representation of the vectors fig 2. DeepSeek limited itself to a text response without visualization fig. 3.

Fig. 2. ChatGPT provided a solution with visualization

Fig. 3. DeepSeek did not build the drawing

Micro-inference: For tasks that require visualization and detailed explanations, the capabilities of different AIs can vary significantly.

Task No 3. Curves of the second order.

Both AIs correctly converted the canonical equation. However, errors were made in calculating the focus coordinates fig. 4, 5.

Fig. 4. Calculating the focal length in DeepSeek

Fig. 5. Drawing for the task made in ChatGPT

Micro-inference: Even with high computational accuracy, artificial intelligence can make individual mathematical errors.

In the course of the study, tasks in linear algebra, vector algebra and analytical geometry were considered. It has been established that both AIs demonstrate high accuracy in standard calculations, but in tasks that require visualization and interpretation, the quality of solutions differs.

3. Conclusion

The use of artificial intelligence in teaching higher mathematics helps to increase the visibility and speed of obtaining solutions. However, the use of AI requires mandatory verification of the results obtained and critical analysis by students.

References
1. Lande D.V. Artificial intelligence in education. – M. : 2022.

2. Goodfellow I., Bengio Y., Courville A. Deep Learning. MIT Press, 2016.