DEEP LEARNING FOR CRITICAL INFRASTRUCTURE SECURITY IN ENERGY AND TRANSPORTATION

This article explores the role of deep learning methodologies in strengthening the security of critical infrastructures in the energy and transportation sectors. It examines both physical and cyber threats, analyzing the applications of Convolutional Neural Networks, Recurrent Neural Networks, Transformers, Autoencoders, and hybrid architectures. The study highlights the advantages of deep learning for anomaly detection, threat forecasting, and real-time decision support, with examples such as SCADA protection in power grids, UAV detection over energy sites, and attack detection in vehicular networks. Comparative evaluation shows that deep learning models significantly surpass traditional rule-based systems, achieving high accuracy in identifying intrusions, equipment failures, and prohibited objects. At the same time, challenges related to data scarcity, model explainability, adversarial robustness, and resource demands are discussed. The findings suggest that integrating explainable AI, federated learning, and multimodal approaches will enhance the resilience of energy and transportation infrastructures, contributing to national security and sustainable development.