Intelligent programming support system: machine learning feat. Fast development of secure programs.

Abstract

The article is devoted to the field of software development. The considered scientific contradiction lies in the fact that, on the one hand, the use of manual labor of a programmer is necessary in this area, and on the other hand, the presence of a human factor negatively affects the safety of the resulting code. To resolve the contradiction, it is proposed to use machine learning, which is traditionally used to solve the problem of classification, regression, search for anomalies, clustering, generalization and search for associations. It is shown that the majority of publications on this solution are of a private nature and do not cover the entire spectrum of possibilities. Various ways of automating the programming process using solutions for the specified machine learning problems are considered and substantiated. The demand for a system that combines such methods is indicated; Also, for the first time, its author’s definition is introduced: «Intelligent Programming Support System – a computer automated system based on artificial intelligence technologies, the purpose of which is to help developers of program code in the interests of reducing and simplifying manual labor, as well as increasing the safety of the final product». A comparative analysis of automation methods based on machine learning is given according to 8 criteria that this intelligent system must meet. The ways of further continuation of the research are indicated.