UDC 004.032.26
ANALYSIS AND RESEARCH OF NEURAL NETWORK TECHNIQUES APPLICATION FOR PROGRAM CODE GENERATION
Yu. L. Leokhin, Dr. Sc. (Tech.), full professor, Vice-Rector for Scientific Work, MTUCI, Moscow, Russia;
orcid.org/0000-0003-3321-4497, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
T. D. Fatkhulin, Ph.D. (Tech.), associate professor, department of MC and IT, MTUCI, Moscow, Russia;
orcid.org/0000-0003-0998-1055, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
M. S. Kozhanov, student, MTUCI, Moscow, Russia;
orcid.org/0009-0007-6567-2819, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The problem of studying the influence of a user-composed prompt on the efficiency of the generated program code is considered. The aim of the work is to develop a methodology that makes it possible to analyze the operation of online neural network services that generate program code and evaluate their effectiveness. The relevance of the work is due to the widespread introduction of neural network techniques in all spheres of human life, including software development. The online neural network services «ChatGPT 3.5», «ChatGPT 4», «YandexGPT 2» and «GigaChat» are considered, which allow generating program code in a specific programming language in question-and-answer form. A methodology is given for determining the effectiveness of neural network tools used to generate program code. The developed methodology is tested by introducing multi-level tasks into neural network services under consideration and analyzing the program code generated by them. In conclusion, the results of the analysis of data obtained during the experiments are presented and conclusions are drawn.
Key words: : neural network techniques, programming, automation, artificial intelligence, program code, advantages, disadvantages, optimization, development.