A computational framework for IoT security integrating deep learning-based semantic algorithms for real-time threat...

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Internet of Things

Background:

• The proliferation of Internet of Things (IoT) networks presents significant security challenges, particularly in real-time threat detection and response.
• Existing security solutions often struggle with the dynamic and complex nature of modern cyber threats targeting IoT ecosystems.

Purpose of the Study:

• To develop and validate a novel hybrid deep learning and semantic reasoning framework for advanced IoT threat intelligence and autonomous security.
• To enhance the capabilities for low-latency identification and response to sophisticated cyber-attacks within IoT environments.

Main Methods:

• Integration of Convolutional Neural Networks (CNNs) for spatial anomaly detection and Recurrent Neural Networks (RNNs) for sequential pattern recognition.
• Implementation of a semantic contextualization layer using knowledge graphs for context-aware threat detection.
• Application of Edge Computing and Real-Time Stream Processing paradigms for efficient, low-latency threat analysis.

Main Results:

• Demonstrated high accuracy, scalability, and adaptability in identifying advanced persistent threats (APTs) and distributed denial-of-service (DDoS) attacks.
• Validated the framework's computational and energy efficiency using the CICIoT 2023 dataset and a custom IoT testbed.
• Achieved low-latency threat identification crucial for real-time response.

Conclusions:

• The proposed hybrid framework effectively bridges deep learning, semantic reasoning, and practical IoT security challenges.
• The research contributes to next-generation autonomous IoT security solutions, emphasizing responsible deployment with privacy and ethical considerations.
• Future work will focus on real-world deployments and adaptive threat intelligence mechanisms.