Local 5G, otherwise known as private networking, has a huge economic potential, with some studies estimating its valuation at $36 billion in 2030, up from $1.1 billion in 2022. Its evolution, local 6G, will predominantly serve highly localized scenarios where wireless connectivity enables new vertical applications and digitalization in fields as diverse as healthcare, manufacturing, or retail. Unlike conventional broadband, these use cases require reliable, resilient, and secure connectivity. This can be achieved with the reflective intelligent surface technology that can reflect or scatter signals already propagating in the environment. However, reflective intelligent surfaces cannot effectively direct the reflected signals without sufficiently accurate information about the wireless environment. This research project focuses on addressing this issue and developing a new neural network-based method for channel representation with a focus on advancing the concept of local 6G. To this end, the project investigates the theoretical foundations of the proposed method, experimental validation, and its implications for local 6G architecture, design, and security. The project contributes to an inter-continental alignment between visions for 6G by engaging the academic research and industry communities involved in pre-standardization activities in the U.S. and Europe. Other broader impacts include advances in workforce development and educational activities, enhancing diversity, and disseminating academic results to the public.<br/><br/>This project advances the concept of private networks by introducing controllable, resilient, and secure local 6G that features novel neural network-based wireless channel representation, a network architecture that integrates reflective intelligent surfaces, and security solutions based on geofencing. The project is divided into three research thrusts, addressing different pillars of local 6G: controllability (Thrust 1), resiliency (Thrust 2), and security (Thrust 3). Thrust 1 develops a novel method of Neural Wireless Channel, based on neural processes, to track, predict, and control the wireless channel. The experimental validation of the method entails capturing a unique dataset that may be used to advance research in artificial intelligence applications to other areas of communications. Thrust 2 develops a novel stochastic framework for analyzing reliability and resiliency in local 6G that comprises reflective surfaces and neural network-based channel representation. Thrust 3 reports on new security vulnerabilities and threats associated with using reflective surfaces and neural network-based channel models and proposes a mechanism based on geofencing to address these security gaps. Overall, the proposed research agenda focuses on the new channel representation methods to advance reflective intelligent surface technology and enable local 6G connectivity.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.