Predicting Dynamics of Transmon Qubit-Cavity Systems with Recurrent Neural Networks

TL;DR

This paper introduces a recurrent neural network approach to efficiently model the complex non-Markovian dynamics of transmon qubit-cavity systems, surpassing traditional master equation solutions in accuracy and computational cost.

Contribution

The study presents a novel neural network-based method to approximate solutions to the Lindblad master equation for quantum systems, capturing non-Markovian effects more efficiently.

Findings

The neural network accurately predicts quantum observables from microscopic dissipative mechanisms.

The model outperforms traditional methods in computational efficiency.

It successfully maps system dynamics from simulated data.

Abstract

Developing accurate and computationally inexpensive models for the dynamics of open-quantum systems is critical in designing new qubit platforms by first understanding their mechanisms of decoherence and dephasing. Current models based on solutions to master equations are not sufficient in capturing the non-Markovian dynamics at play and suffer from large computational costs. Here, we present a method of overcoming this by using a recurrent neural network to obtain effective solutions to the Lindblad master equation for a coupled transmon qubit-cavity system. We present the training and testing performance of the model trained a simulated dataset and demonstrate its ability to map microscopic dissipative mechanisms to quantum observables.