Today’s buildings are complex cyber-physical systems, with a variety of sensors emitting data based on the building usage and controllers managing their heating, ventilation, air conditioning, lighting, major appliances, and access control systems. The impressive advances we have been witnessing in sensing and control technologies (under the general Internet of Things banner) coupled together with new machine-learning methods for modeling and predicting the behaviors of complex dynamic systems offer a timely and compelling opportunity for optimizing this sense-and-control loop, and consequently curtailing the substantial energy impact of buildings on the environment and substantially reducing their operation costs, while also improving the comfort of their occupants.