Abstract:

The swift IoT devices has transformed industries, but it has also made networks more vulnerable to a wide range of cyber threats. This study develops deep learning-based models to identify malicious activity in IoT ecosystems, utilizing a combination of Generative Adversarial Networks (GANs), Capsule Networks, and Multi-Layer Perceptrons (MLPs). GANs are employed to address data imbalance by generating synthetic IoT data, including uncommon attack scenarios. Capsule Networks are used to detect intricate attack patterns by analyzing complex feature relationships. Finally, the MLP classifier leverages these rich representations to accurately differentiate between benign and malicious behaviors. Experimental findings highlight the efficacy of the suggested model in enhancing the security of IoT networks.

Description:

.