Text this: Viewpoint Selection for the Efficient Teleoperation of a Robot Arm Using Reinforcement Learning