Resource-efficient high-dimensional subspace teleportation with a quantum autoencoder.

Area of Science:

• Quantum Information Science
• Quantum Machine Learning
• Integrated Photonics

Background:

• Quantum autoencoders offer efficient quantum data compression.
• Quantum teleportation is crucial for quantum communication and computing.
• High-dimensional quantum systems, like qutrits, offer enhanced capabilities but require efficient resource management.

Purpose of the Study:

• To propose and demonstrate the use of quantum autoencoders for resource-efficient quantum teleportation of subspaces in high-dimensional systems.
• To achieve the first demonstration of qutrit teleportation using quantum autoencoders on an integrated photonic platform.
• To explore the application of unsupervised machine learning for training on-chip autoencoders for state compression and teleportation.

Main Methods:

• Utilizing a compress-teleport-decompress protocol with a quantum autoencoder.
• Implementing the scheme on an integrated photonic platform for scalability.
• Applying unsupervised machine learning to train the on-chip autoencoder for dimensionality reduction and state recovery.
• Teleporting unknown quantum states encoded within high-dimensional subspaces.

Main Results:

• Demonstrated chip-to-chip quantum teleportation of qutrit states using an autoencoder.
• Achieved high fidelity (~0.971) in decompressing unknown states.
• Obtained a total teleportation fidelity of ~0.894.
• Showcased the ability to compress dimensionality by erasing redundant information.

Conclusions:

• Quantum autoencoders effectively reduce resource costs for high-dimensional quantum teleportation.
• The developed scheme provides a scalable platform for future quantum computing and networking.
• Subspace encodings, enabled by this technique, offer improved noise robustness and coherence.
• This work lays the foundation for integrating machine learning techniques into quantum systems.