Tools
Tools
Murukutla, Hemanth (2023) Detecting IOT Attacks Using Artificial intelligence. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (760kB) | Preview |
Preview |
PDF (Configuration manual)
Download (831kB) | Preview |
Abstract
The rapid expansion and significant significance of IIoT networks have led to a concerning escalation in cyber vulnerabilities, thereby demanding the creation of more advanced detection methodologies. Traditional cybersecurity techniques have proven ineffective in protecting these complex systems, emphasizing the necessity for novel and advanced approaches. The objective of the research is to assess the efficiency of machine learning models, with a specific focus of the Variational Autoencoder (VAE), Long Short-term Memory Model (LSTM), Recurrent Neural Network (RNN) and the Gated Recurrent Unit (GRU), in recognizing cyber threats within IIoT networks. The Edge IoT dataset, which is an extensive set of network logs gathered from several common IIoT settings, was used as the basis for this investigation. This set of data was utilized during the training, testing, and evaluation of the various machine learning models.. The research study validates the usefulness of the GRU model, exhibiting an amazing detection accuracy rate of 97%. These findings demonstrate the great potential of applying the GRU model for detecting cyber risks in IIoT networks. This research contributes to the larger mission of improving cybersecurity by fortifying our interconnected industrial systems against cyber-attacks.
Actions (login required)
 |
View Item |