Deterministic generation of nonclassical mechanical states in cavity optomechanics via reinforcement learning

TL;DR

This paper introduces a reinforcement learning-based method to deterministically generate nonclassical mechanical states in cavity optomechanics, achieving high-fidelity phononic and entangled states for quantum technology applications.

Contribution

It presents a novel reinforcement learning approach to optimize driving pulses for deterministic nonclassical state generation in cavity optomechanics.

Findings

High-fidelity phononic Fock states achieved

Generation of superposed Fock states demonstrated

Creation of two-mode entangled states shown

Abstract

Nonclassical mechanical states, as vital quantum resources for exploring macroscopic quantum behavior, have wide applications in the study of the fundamental quantum mechanics and modern quantum technology. In this work, we propose a scheme for deterministically generating non-classical mechanical states in cavity optomechanical systems. By working in the eigen-representation of the nonlinear optomechanical systems, we identify the carrier-wave resonance conditions and seek optimal driving pulses for state preparations. Concretely, we employ the reinforcement learning method to optimize the pulsed driving fields, effectively suppressing the undesired transitions induced by both the pulsed driving fields and dissipations. This approach enables the high-fidelity preparation of phononic Fock states and superposed Fock states in the single-resonator optomechanical systems, as well as two-mode entangled states in the two-resonator optomechanical systems. The statistical properties of the generated states are also examined. Our results create an opportunity for quantum state engineering in quantum optics and quantum information science via reinforcement learning.