Abstract: We propose a frequency-domain state representation to improve the performance and reduce the computation and data requirements of reinforcement learning. This approach is tailored to ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results