Dual-Capability Machine Learning Models for Quantum Hamiltonian Parameter Estimation and Dynamics Prediction.

Area of Science:

• Quantum computing
• Quantum many-body systems
• Machine learning applications

Background:

• Quantum system data accessibility has increased due to hardware and simulation advancements.
• Accurate prediction of quantum Hamiltonian dynamics and parameter identification are vital for quantum simulations, error correction, and control.

Purpose of the Study:

• To develop a machine learning model capable of deducing time-dependent Hamiltonian parameters from local observables.
• To enable prediction of observable evolution based on Hamiltonian parameters.
• To enhance quantum computing tasks like parameter estimation and control.

Main Methods:

• A novel machine learning model was developed with dual capabilities for Hamiltonian parameter deduction and observable evolution prediction.
• The model's performance was validated through theoretical simulations.
• Experimental validation was conducted on nuclear magnetic resonance and superconducting quantum computers.

Main Results:

• The model accurately predicted the dynamics of local observables on a nuclear magnetic resonance quantum computer.
• The model successfully inferred unknown Hamiltonian parameters on a superconducting quantum computer.
• The dual-capability model demonstrated robust performance across various scenarios.

Conclusions:

• The developed machine learning model effectively deduces Hamiltonian parameters and predicts quantum system dynamics.
• This approach significantly enhances capabilities in quantum parameter estimation, noise characterization, and quantum control optimization.
• The model's successful experimental validation paves the way for broader applications in quantum information science.