Experimental Property Reconstruction in a Photonic Quantum Extreme Learning Machine

Area of Science:

• Quantum Information Science
• Machine Learning Applications
• Photonics

Background:

• Characterizing quantum states is crucial for quantum information processing.
• Machine learning integration into experimental platforms offers new solutions.
• Photonics provides a robust platform for quantum experiments.

Purpose of the Study:

• To implement a resource-efficient and accurate method for characterizing photon polarization states.
• To leverage quantum extreme learning machines (QELM) in a photonic setup.
• To demonstrate robustness against experimental imperfections.

Main Methods:

• Implementation of a QELM using a photonic platform.
• Utilizing coined quantum walks of high-dimensional photonic orbital angular momentum for reservoir dynamics.
• Performing projective measurements over a fixed basis.

Main Results:

• Achieved resource-efficient and accurate characterization of photon polarization states.
• Demonstrated that reconstruction of unknown polarization states does not require detailed characterization of the measurement apparatus.
• Showcased robustness to experimental imperfections.

Conclusions:

• The developed QELM photonic platform offers a promising route for resource-economic quantum state characterization.
• This approach simplifies experimental requirements and enhances reliability.
• Highlights the potential of integrating machine learning with quantum experiments for practical applications.