Abstract: A method for controlling a robot device. The method includes performing an initial training of an actor neural network by imitation learning of demonstrations, controlling the robot device by the initially trained actor neural network to generate multiple trajectories, wherein each trajectory comprises a sequence of actions selected by the initially actor neural network in a sequence of states, and observing the return for each of the selected actions, performing an initial training of a critic neural network by supervised learning, wherein the critic neural network is trained to determine the observed returns of the actions selected by the initially actor neural network, training the actor neural network and the critic neural network by reinforcement learning starting from the initially trained actor neural network and the initially trained critic neural network and controlling the robot device by the trained actor neural network and trained critic neural network.

Abstract: A machine control system including a first data processing device having a first central processing unit, which implements a control agent, the first data processing unit communicating control commands to the machine and receiving process and environmental data from the machine, which contain state information about the state of the machine and/or an environment of the machine. A second data processing device receives the state information from the first data processing device and implements a copy of the control agent. The second data processing device ascertains, with the aid of one or multiple graphics processing units and/or one or multiple tensor processing units, an update of the copy of the control agent by reinforcement learning, using the state information, and communicates the update to the first data processing device.

Abstract: A method for controlling a robot device. The method includes performing an initial training of an actor neural network by imitation learning of demonstrations, controlling the robot device by the initially trained actor neural network to generate multiple trajectories, wherein each trajectory comprises a sequence of actions selected by the initially actor neural network in a sequence of states, and observing the return for each of the selected actions, performing an initial training of a critic neural network by supervised learning, wherein the critic neural network is trained to determine the observed returns of the actions selected by the initially actor neural network, training the actor neural network and the critic neural network by reinforcement learning starting from the initially trained actor neural network and the initially trained critic neural network and controlling the robot device by the trained actor neural network and trained critic neural network.