Lorenzo Cardarelli (FZ Jülich and RWTH Aachen)

Quantum Error Correction with Quantum Autoencoders

Active quantum error correction is a central ingredient to achieve robust quantum computation. Our work investigates the potential of quantum machine learning for error suppression in quantum memories, to achieve logical qubit lifetime extension. We show that quantum autoencoders can be trained to learn optimal strategies for active detection and correction of errors, including qubit losses, and highlight that the denoising capabilities are not limited to the protection of specific states but extend to the entire logical codespace. We show that quantum autoencoders can be used to discover optimal logical encodings for the underlying noise. Lastly, we find that beneficial error mitigation is attainable also in realistic scenarios of moderately noisy autoencoders.