Deep Reinforcement Learning Control of Quantum Cartpoles
This work demonstrates the applicability of deep learning to continuous-space quantum control, potentially benefiting researchers in quantum physics and control engineering, though it appears incremental as it adapts existing methods to a new domain.
The authors tackled the problem of stabilizing a quantum particle in an unstable potential, analogous to a classical cartpole, using deep reinforcement learning with measurement and feedback, achieving performance comparable to or better than standard control theory methods.
We generalize a standard benchmark of reinforcement learning, the classical cartpole balancing problem, to the quantum regime by stabilizing a particle in an unstable potential through measurement and feedback. We use state-of-the-art deep reinforcement learning to stabilize a quantum cartpole and find that our deep learning approach performs comparably to or better than other strategies in standard control theory. Our approach also applies to measurement-feedback cooling of quantum oscillators, showing the applicability of deep learning to general continuous-space quantum control.