Defesa cibernética: o emprego de sistemas de detecção de intrusão com apoio de aprendizado de máquina para otimizar a detecção de ataques

Abstract

ABSTRACT CYBER DEFENSE: THE USE OF INTRUSION DETECTION SYSTEMS SUPPORTED BY MACHINE LEARNING TO OPTIMIZE ATTACK DETECTION. Cyberattacks have expanded in recent years, causing devastating impacts. As a result, state concerns over cybersecurity have increased, leading to the adoption of protective measures aimed at mitigating existing vulnerabilities. These measures have encouraged research and the implementation of technologies focused on protecting information and critical infrastructures. Intrusion Detection Systems (IDS) are essential tools for monitoring and preventing cyberattacks, with the goal of making cyberspace safer. Artificial intelligence solutions, specifically those based on machine learning, have been increasingly used to automate attack detection and address the complex challenges of cybersecurity.This work aims to analyze the contribution of machine learning in the implementation of IDS systems. To achieve this goal, state-of-the-art research employing machine learning and deep learning techniques in the implementation of intrusion detection systems was conducted. Based on the analyzed research, it was concluded that machine learning and deep learning offer promising paths for intrusion detection, with supervised techniques such as SVM and ANN already demonstrating significant success, while deep learning approaches, such as CNN and RNN, offer even greater accuracy and the ability to detect complex patterns.