The widespread adoption of the Internet of Things (IoT) forms a critical foundation for enabling applications in healthcare, transportation, and industrial automation. However, the ultra-dense deployment of IoT devices and their need to transmit sensitive data raise significant challenges for efficient and secure communication. Conventional cryptographic methods are often too computationally intensive for resource-constrained IoT devices. This project explores a lightweight and non-cryptographic framework that secures wireless communication by leveraging the randomness of physical wireless channels, grounded in information-theoretic principles of Physical Layer Security (PLS). To address challenges in dense networks where channel correlation among users is high, the project integrates Intelligent Reflecting Surfaces (IRS), passive devices capable of reconfiguring wireless signal paths, into the system design to improve both security and energy efficiency. In addition to its technical contributions, the project supports national workforce development by providing interdisciplinary research training, enhancing cybersecurity education, and engaging students across multiple institutions. This project investigates a learning-based framework to enhance Physical Layer Security (PLS) and energy efficiency in ultra-dense IoT networks using Intelligent Reflecting Surfaces (IRS). By dynamically adjusting the IRS configuration based on relational information among legitimate users and