A carregar...
Publicação
IDS model for identifying Cyber Threats: Applying the novel Kolmogorov Arnold Neural Networks to the contemporary cyber-attack datasets, UNSW-NB15 and CICIDS2017
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1021.09 KB | Adobe PDF |
Orientador(es)
Resumo(s)
The Kolmogorov Arnold Neural Network variant of neural networks is a novel discovery in
the field of Deep Learning introduced earlier this year. This algorithm is being hypothesized
as a strong contender for superiority in performance as well as explainability in comparison
to other artificial neural networks already in use. By leveraging the existing body of
literature, this thesis aims to explore the performance of this novelty in deep learning when
applied to Intrusion Detection Systems. Through its application onto two up-to-date
datasets, the UNSW-NB15 and CICIDS2017, this thesis will provide ground for comparison
between the previous state-of-the-art models and Kolmogorov Arnold Neural Networks,
effectively looking to understand if this novelty can surpass previous models within the field
of deep learning applied to cybersecurity.
Descrição
Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Business Intelligence
Palavras-chave
Kolmogorov Arnold Neural Network Intrusion Detection Systems Deep Learning Performance Cybersecurity SDG 9 - Industry, innovation and infrastructure