Paper Key : IRJ************231
Author: Sharayu Anil Patil
Date Published: 11 Jul 2024
Abstract
This method sounds promising for enhancing intrusion detection in cyber-physical systems. Leveraging deep learning particularly generative adversarial networks (GANs), can indeed offer advantages in identifying cybersecurity vulnerabilities and breaches.By contrasting unsupervised and deep learning-based discriminative approaches, you're likely exploring a range of methodologies to address the limitations of traditional intrusion detection systems. GANs, in particular, are known for their ability to generate synthetic data that closely resembles real data, which could be highly beneficial in detecting novel types of intrusions.The reported performance increase of 95% to 97% in terms of accuracy, reliability, and efficiency is impressive. Achieving such gains is crucial in the realm of cybersecurity where the detection of attacks needs to be both timely and precise. Additionally, setting the dropout value to 0.2 and the epoch value to 25 seems to have contributed to achieving these results, indicating the importance of hyperparameter tuning in deep learning models.KEY WORDS - Cyber security, Internet of Things, Intrusion detection system (IDS)Anomaly detection, Security attacks, Deep learning, Network Security
DOI Requested