This repository provides the videos and implementation codes that accompany the paper: Deep Reinforcement Learning Control of Quantum Cartpoles at https://arxiv.org/abs/1910.09200
The videos demonstrate how the deep reinforcement learning controller controls the particles by showing the wavefunctions and the probability densities, and they are compared with the other best-performing strategies. In the videos, a short wavelength of the wavefunction corresponds to a large velocity.
The deep reinforcement learning controller uses distribution moments of the controlled state as its input, i.e. the one with the highest performance shown in the paper.