Deep Reinforcement Learning Control of Quantum Cartpoles
Deep Reinforcement Learning Control of Quantum Cartpoles
复制标题
DOI：
10.1103/physrevlett.125.100401
复制
发表时间：
2019-10
影响因子：
8.6
通讯作者：
Zhikang T. Wang;Yuto Ashida;Masahito Ueda
中科院分区：
文献类型：
--
作者：
Zhikang T. Wang;Yuto Ashida;Masahito Ueda
我们将强化学习的标准基准，经典的cartpole平衡问题，通过测量和反馈将粒子稳定在不稳定的势中，从而将其推广到量子领域。我们使用最先进的深度强化学习来稳定量子cartpole，并发现我们的深度学习方法比标准控制理论中的其他策略表现更好。我们的方法也适用于量子振荡器的测量反馈冷却，显示了深度学习对一般连续空间量子控制的适用性。
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.