Reinforcement Learning Meets Quantum Control - Artificially Intelligent Maxwell's Demon

Feedback control of quantum systems is of fundamental importance for practical applications in various contexts, ranging from quantum computation to quantum error correction and quantum metrology. However, deriving optimal feedback control strategies is highly challenging, particularly when it involves the optimal control of open quantum systems, the stochastic nature of quantum measurement, and the inclusion of policies that maximize a long-term time- and trajectory-averaged goal.

In our work, we employ a reinforcement learning approach to automate and capture the role of a quantum Maxwell’s demon: a neural network takes the literal role of discovering optimal control-feedback strategies in qubit-based quantum systems that maximize the tradeoff between measurement-powered cooling and measurement efficiency. We explore different operational regimes based on the ordering between thermalization, measurement, and unitary feedback timescales, finding different and highly non-intuitive, yet interpretable, strategies.