Regresar

An Implementation of an Algorithm for Information Theft Using Deep Learning Techniques: An Ethical Hacking Event

Abstract:

Ethical hacking is modern data protection and prevention technique proper to prevent information losses. Under this precept, this study aims to design and implement an algorithm capable of circumventing preventive security to perpetuate an attack to steal information from the victim. The algorithm programmed through Python, MatLab, and R, automates the export of a data repository saved in a.CSV file, which contains all the methods to prevent an attack. In this way, when the algorithm is executed, it is trained to disable all detection mechanisms. The algorithm uses the structure of a neural network based on Deep Learning, Bayesian networks, a mathematical model, and a Rectified Linear Activation function or ReLU. It was configured to execute the algorithm on each interpreted output data. The proofs of concept were developed in a controlled virtual network environment, using a Linux-based server. The results were validated based on resource management in both processing and memory, using monitoring standards and averaging curves. Thus, we demonstrate the successful unauthorized access to a remote server to steal all the victim’s information and that the algorithm was undetectable and efficient.