Learning Temporal Quantum Tomography

Area of Science:

• Quantum Information Science
• Machine Learning
• Quantum Computing

Background:

• Characterizing quantum states (quantum state tomography) is crucial for quantum device development but resource-intensive.
• Existing tomography methods are unsuitable for quantum devices with temporal processing.
• Temporal processing in quantum devices presents unique challenges for state verification.

Purpose of the Study:

• To develop a practical and approximate tomography method for quantum devices with temporal processing.
• To introduce a novel machine learning framework for analyzing quantum temporal dynamics.
• To propose a metric for evaluating the temporal processing capabilities of quantum devices.

Main Methods:

• Utilized a recurrent machine learning framework based on a quantum reservoir system.
• Employed repeated quantum interactions between the quantum reservoir and a stream of quantum states.
• Trained a recurrent relation between quantum channels using measurement data and linear readout.

Main Results:

• Demonstrated the efficacy of the developed algorithms on representative quantum learning tasks.
• Successfully adapted machine learning for tomography in a temporal quantum processing scenario.
• Proposed a quantum memory capacity metric to assess temporal processing ability.

Conclusions:

• The recurrent machine learning approach offers a practical solution for quantum state tomography in temporal devices.
• This method advances the characterization of quantum states in dynamic quantum systems.
• The proposed quantum memory capacity provides a valuable tool for evaluating near-term quantum devices.