H.B. Keller Colloquium

Our electric grids are undergoing changes in both form and function, where renewable resources and new devices are creating systems that are more distributed, dynamic and uncertain. Modern AI and machine learning tools have the potential to transform the operation of these new energy systems. However, such algorithms typically do not provide guarantees about stability or safety, making them difficult to implement in practice. In this talk, I will describe how to bridge these gaps. I will show how structured neural networks can leverage advances in AI and provide formal guarantees such as system stability and hard constraint satisfaction. Then I will close the loop by discussing the challenge of supplying power to data centers and how it could be done efficiently and fairly, and how AI-based algorithms can be used in this respect.