Time Series Prediction of Open Quantum System Dynamics by Transformer Neural Networks.

Area of Science:

• Quantum Information Science
• Computational Physics
• Machine Learning

Background:

• Simulating open quantum systems is vital for quantum information science.
• Exact numerical solutions for the Lindblad master equation are computationally intensive.
• Machine learning offers a promising approach for simulating quantum dynamics.

Purpose of the Study:

• To develop a computationally efficient deep learning model for predicting the dynamics of open quantum systems.
• To utilize time series prediction (TSP) with Transformer neural networks for forecasting quantum system evolution.
• To provide a scalable and practical method for analyzing open quantum system dynamics.

Main Methods:

• Developed a deep learning model using Transformer neural networks for time series prediction (TSP).
• Employed the positive operator-valued measure (POVM) approach to transform density matrices into probability distributions.
• Trained the TSP model to capture historical patterns and predict future system behavior.

Main Results:

• Achieved high-fidelity predictions of system evolution trajectories in both short- and long-term scenarios.
• Demonstrated robust generalization capabilities across different initial states and coupling strengths.
• Successfully predicted the steady-state behavior of the open quantum system.

Conclusions:

• The proposed deep learning TSP model offers a practical and scalable solution for simulating open quantum system dynamics.
• This method provides an efficient alternative to traditional computationally expensive numerical techniques.
• The model's accuracy and generalization ability highlight its potential for advancing quantum information science research.