Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ Repositorio Document...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Repositorio Documental UMNG
Bachelor thesis . 2021
License: CC BY NC ND
versions View all 1 versions
addClaim

Reconocimiento de ataques cibernéticos utilizando aprendizaje profundo

Authors: Mendoza Molina, Leidy Daniela;

Reconocimiento de ataques cibernéticos utilizando aprendizaje profundo

Abstract

La Seguridad Informática es un concepto relacionado con la importancia de proteger la confiabilidad, integridad y disponibilidad (CIA) de los diferentes dispositivos tecnológicos que puedan ser usados por una empresa o un usuario. Busca que las personas tengan confianza en que sus datos personales están seguros, que no existan riesgos de ser alterados, y que estén por fuera del acceso de terceros con fines maliciosos. Sin embargo, a través de los años, así como han avanzado las tecnologías también han avanzado las herramientas para atacar la seguridad informática. Dentro de la vulnerabilidad cibernética se encuentran los siguientes ataques: ataque de denegación de servicio (DoS) y ataque de denegación de servicio distribuido (DDoS). Estos ataques buscan detener un servicio con el fin de que muchas personas dejen de usar el mismo o para destruir programas de un sistema informático, esto por medio de tráfico malicioso. El propósito de este proyecto es utilizar inteligencia artificial, específicamente modelos de aprendizaje profundo (DL: Deep learning), para poder identificar diferentes tipos de ataques cibernéticos. Para cumplir con este propósito se seleccionan dos estrategias de diseño de clasificación de imágenes basadas en DL: búsqueda automática de hiperparámetros (con AutoML), y transferencia de aprendizaje con modelos preentrenados de CNN y aplicando búsqueda manual de hiperparámetros. Específicamente, se escogen tres modelos pre-entrenados para transferencia de aprendizaje: VGG16, ResNet 152 y Xception, con hiperparámetros de profundidad de la red transferida, y optimizadores (RMSProp, Adamax, y SGD). En total, se entrenaron 45 modelos de aprendizaje profundo con el conjunto de imágenes denominado CICDDoS2019 para la clasificación de ataques cibernéticos. Information Security is a concept related to the importance of protecting the reliability, integrity and availability (CIA) of the different technological devices that can be used by a company or a user. It seeks that people have confidence that their personal data is safe, that there is no risk of being altered, and that it is out of the access of third parties for malicious purposes. However, over the years, as technologies have advanced, so have the tools to attack computer security. Cyber vulnerability includes the following attacks: denial of service (DoS) and distributed denial of service (DDoS) attacks. These attacks seek to stop a service in order to stop many people from using it or to destroy programs in a computer system through malicious traffic. The purpose of this project is to use artificial intelligence, specifically deep learning (DL) models, to identify different types of cyber-attacks. To fulfill this purpose, two DL-based image classification design strategies are selected: automatic hyperparameter search (with AutoML), and transfer learning with pretrained CNN models and applying manual hyperparameter search. Specifically, three pre-trained models are chosen for transfer learning: VGG16, ResNet 152 and Xception, with the following hyperparameters: depth of the network and optimizers (RMSProp, Adamax, and SGD). In total, 45 deep learning models were trained with the image dataset named CICDDDoS2019 for cyber-attack classification. Pregrado

Country
Colombia
Related Organizations
Keywords

aprendizaje profundo, modelos pre-entrenados, APRENDIZAJE PROFUNDO (APRENDIZAJE AUTOMATICO), cybersecurity, deep learning, ciberseguridad, transferencia de aprendizaje, pre-trained models, SEGURIDAD EN COMPUTADORES, transfer learning, CIBERESPACIO, DoS, DDoS, AutoML, CNN

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
0
Average
Average
Average
Green
Related to Research communities