Task-Optimal Exploration in Linear Dynamical Systems

                Andrew Wagenmaker 1 Max Simchowitz 2 Kevin Jamieson 1

             Abstract                  tainty about the environment, a naive strategy might be to
   Exploration in unknown environments is a fun-          explore the environment until it is uniformly understood
   damental problem in reinforcement learning and         (system identification), and then devise a plan to complete
   control. In this work, we study task-guid ...