Brain-Box Quantum Autoencoder for entangled state production

The Quantum Autoencoder (QAE) is a quantum neural network. In presence of harmful noise channels, the QAE can be trained to generate high-fidelity Greenberger-Horne-Zeilinger states. We identify a few failure mechanisms in the process and we show a large improvement in the noise tolerance by using a multi-qubit structure called "brain box" (BB).