Data compression for quantum machine learning

The advent of noisy-intermediate scale quantum computers has introduced the exciting possibility achieving speedups in machine learning tasks. These devices, however, are composed a small number qubits, and can faithfully run only short circuits. This puts many proposed approaches for beyond currently available devices. We address problem efficiently compressing loading classical data use on computer. Our methods allow both required qubits depth circuit to be tuned. achieve this by using correspondence between matrix-product states circuits, further propose hardware-efficient approach, which we benchmark Fashion-MNIST dataset. Finally, demonstrate that based classifier competitive accuracy with current tensor 11 qubits.