An innovative hybrid ANN inception model for detecting and preventing man-in-the-middle attacks in smart networks for shielding sensitive data

Man-in-the-Middle (MitM) attacks represent a significant cybersecurity challenge, particularly within the rapidly growing domain of smart networks and Internet of Things (IoT) environments. These attacks involve an adversary intercepting and potentially altering the communication between two parties who believe they are communicating directly with each other. As the deployment of smart devices and autonomous systems increases, the risk of MitM attacks escalates, given the sensitive nature of the data being transmitted and the critical functions these systems perform. This paper presents a novel approach to detecting and preventing MitM attacks by leveraging a Hybrid ANN-Inception Model (HAINet). The proposed model integrates the dynamic feature extraction capabilities of Inception modules with the flexibility of Artificial Neural Networks (ANNs) to identify and mitigate these sophisticated cyber threats. Using the Smart Home Intrusion Detection Dataset from Kaggle, the methodology involves advanced data preprocessing, feature extraction via AutoEncoders, and optimization through hybrid learning algorithms. The HAINet model executed in Jupyter Notebook and achieved an outstanding accuracy of 98.69%, outperforming traditional models in both detection accuracy and robustness across various network conditions. The results demonstrate the model's efficacy in shielding sensitive data from interception and manipulation, making it a valuable tool for enhancing security in smart networks.