Reinforcement Learning Enhancing Entanglement for Two-Photon-Driven Rabi Model

TL;DR

This paper introduces a reinforcement learning-based control scheme to enhance entanglement in a two-photon-driven Rabi model, demonstrating robustness and generalization in quantum resource management.

Contribution

It presents a novel reinforcement learning approach to optimize entanglement in a Rabi model, showing improved control and robustness against dissipation.

Findings

Entanglement can be enhanced across various parameter regimes.

The control scheme is robust against dissipation effects.

The method generalizes to different controlled systems.

Abstract

A control scheme is proposed that leverages reinforcement learning to enhance entanglement by modulating the two-photon-driven amplitude in a Rabi model. The quantum phase diagram versus the amplitude of the two-photon process and the coupling between the cavity field and the atom in the Rabi model, is indicated by the energy spectrum of the hybrid system, the witness of entanglement, second order correlation, and negativity of Wigner function. From a dynamical perspective, the behavior of entanglement can reflect the phase transition and the reinforcement learning agent is employed to produce temporal sequences of control pulses to enhance the entanglement in the presence of dissipation. The entanglement can be enhanced in different parameter regimes and the control scheme exhibits robustness against dissipation. The replaceability of the controlled system and the reinforcement learning module demonstrates the generalization of this scheme. This research paves the way of positively enhancing quantum resources in non-equilibrium systems.