AUTOMATED PENETRATION TESTING METHOD USING DEEP MACHINE LEARNING TECHNOLOGY

The article developed a method for automated penetration testing using deep machine learning technology. main purpose of the development is to improve security computer systems. To achieve this goal, analysis existing methods was carried out and their disadvantages were identified. They are mainly related subjectivity assessments in case manual testing. In cases testing, most authors confirm fact that there no unified effective solution procedures used. This contradiction resolved intelligent analysis. It proposed be based on reinforcement study Shadov system's ability collect factual data designing attack trees, as well Mulval platform generating trees. A forming matrix cyber intrusions tool has been developed. Deep Q - Lerning Network improved analyzing intrusion finding optimal trajectory. study, according method, reward scores assigned each node, CVSS rating, made it possible shrink trees identify an with greater likelihood occurring. comparative out. practical possibility system revealed.