Методи виявлення кіберзагроз у хмарних середовищах на основі автоенкодерів

Abstract

This work examines methods for detecting cyber threats in cloud environments based on autoencoders. The rapid spread of cloud technologies today is leading to an increasing number of cyber threats aimed at cloud infrastructure. Traditional signature-based attack detection methods are limited in identifying new or modified threats, which could be solved by using advanced machine learning methods. The study considers the features of using autoencoders as deep learning models for anomaly detection in cloud environments. It is concluded that the use of autoencoders is a promising direction for improving the efficiency of cyber threat detection in cloud environments