Redes neurais recorrentes para geração de senhas em ataques de força bruta baseado em dicionário

• Autor(a) principal: Ramos, Marlon Brendo
• Data de Publicação: 2022
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFU
• Texto Completo: https://repositorio.ufu.br/handle/123456789/36723
• topic: Redes neurais recorrentes; Arquitetura LSTM; Aprendizado de máquina; Geração de senhas; Ataque de dicionário; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

Cyberattack is one of the major problems faced by contemporary society and, notoriously, impacts several business and personal branches. One of the most reported incidents refers to scanning, in which the attacker collects various information from various ports and protocols and makes his attack plan possible. One of the most commonly used types of attacks that stem from scanning is the dictionary-based brute-force attack. Traditionally, brute-force techniques use leaked dictionaries or ones generated through combinations of characters or personal profiles. Thus, the success of this type of attack is linked to the quality and variety of the passwords obtained from this medium. The way this type of attack is carried out, essentially based on these dictionaries, makes it impossible for traditional tech niques to capture semantic relationships between data, such as, for example, the formation of patterns in password scripts. In contrast, the use of neural networks has proven to be an advantageous way to recognize complex or numerous patterns, standing out through methods of dynamic nature with the possibility of realizing predictive systems, showing continuous capacity for improvement and adaptation to changes in patterns. Thus, the main objective of this work is the application of recurrent neural networks, through the LSTM architecture, for the generation and prediction of passwords based on dictionaries used in brute force attacks and, in this way, analyze the feasibility and balance of its use.