A Novel Hybrid Deep Learning Model for Attack Detection in IoT Environment: Convolutional Neural Network with...

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Network Security

Background:

• The proliferation of Internet of Things (IoT) devices has increased network connectivity and accessibility.
• This heightened interconnectivity introduces significant new security vulnerabilities and threat levels.
• Robust anomaly detection systems are crucial for securing modern IoT environments.

Purpose of the Study:

• To present a novel deep learning-based security method for mitigating threats in Internet of Things (IoT) networks.
• To develop and evaluate a hybrid Convolutional Neural Network (CNN) and Transformer model for real-time attack detection and classification.
• To assess the model's performance in terms of accuracy, efficiency, and computational complexity.

Main Methods:

• A hybrid CNN-Transformer deep learning model was designed for analyzing network traffic patterns.
• The model was trained and validated using the comprehensive CIC-IoT-2023 dataset, encompassing 33 diverse IoT threat types.
• Performance metrics including precision, recall, F1 score, accuracy, and loss were meticulously evaluated.

Main Results:

• The proposed hybrid CNN-Transformer model achieved exceptional performance metrics: 99.96% precision, 99.96% recall, 99.96% F1 score, and 99.97% accuracy.
• The model demonstrated a low loss of 0.0123, indicating highly effective learning and prediction.
• Analysis confirmed minimized computational complexity and resource consumption while maintaining superior attack detection capabilities.

Conclusions:

• The developed deep learning approach offers a highly effective and efficient solution for securing Internet of Things networks against sophisticated cyber threats.
• The hybrid CNN-Transformer model provides a robust framework for real-time anomaly detection and classification in complex IoT ecosystems.
• This research contributes a significant advancement in IoT security, balancing high accuracy with computational efficiency.