Quantum computers are becoming capable of simulating increasingly complex quantum materials. Identifying the different quantum phases of these materials is essential for understanding their properties, but doing so with classical methods requires a prohibitively large number of measurements.
In this work, Markus K. Hoffmann, Leon C. Sander, Colin Scarato, Christoph Hellings, Johannes Knörzer, Michael J. Hartmann, and Petr Zapletal introduce a hybrid quantum-classical neural network for recognizing phases of matter that combines a trainable quantum algorithm with a classical neural network. By jointly optimizing the quantum and classical components, the hybrid network learns to identify features that distinguish different quantum phases. The authors show that this approach significantly reduces the number of measurements required for both training and inference compared with a state-of-the-art classical approach based on direct measurements.
Because the hybrid neural network requires only a relatively simple quantum algorithm, it can be readily implemented on existing quantum hardware. This is demonstrated experimentally on a superconducting quantum processor in the companion work by Colin Scarato, Johannes Knörzer, Markus K. Hoffmann, Leon C. Sander, Luca Hofele, Shengpu Wang, Kilian Hanke, Ashay Sathe, Dominic Hagmann, Alexander Flasby, Michael J. Hartmann, Petr Zapletal, Andreas Wallraff, and Christoph Hellings, highlighting a practical route toward understanding complex quantum states.