Designing a neuro-symbolic dual-model architecture for explainable and resilient intrusion detection in IoT networks

The Internet of Things (IoT) is rapidly evolving into a vast ecosystem of interconnected devices that serve diverse domains, including smart homes, healthcare, and gaming. However, the increasing complexity of device behavior, growing cybersecurity threats, and the need for real-time personalized services pose significant challenges in design, performance, and user trust. Traditional AI approaches, while powerful at pattern recognition, often lack interpretability and symbolic reasoning capabilities crucial for sensitive, adaptive consumer environments. In this work, we address these challenges by proposing a hybrid neuro-symbolic AI framework for cyber threat analysis in consumer electronic platforms. Leveraging the NF-BoT-IoT-V2 dataset, we implemented and evaluated both 1D Convolutional Neural Networks (CNN) and Artificial Neural Networks (ANN) to classify cyberattack behaviors with high fidelity. Our models achieved over 98% accuracy, demonstrating the effectiveness of neural architectures in real-world attack detection. To enhance trust and interpretability, we integrated Explainable AI techniques, including SHAP and LIME, to provide feature-level insights into model predictions. Furthermore, we outline the integration of symbolic reasoning and real-world use-case mappings in smart home automation, IoT networks, and gaming environments. This study presents a scalable and interpretable neuro-symbolic approach that advances the intelligence, security, and personalization capabilities of modern consumer electronics.