University of Washington researchers use AI and quantum computing to accelerate materials development

University of Washington researchers used artificial intelligence and quantum computing to accelerate identification of quantum materials. Quantum materials have properties governed by quantum mechanics, including superconductivity, entanglement and unusual magnetism. These properties stem from repeating atomic patterns in crystals and can be controlled at human scales for use in quantum computing and energy-efficient electronics. New materials require intense modeling because properties emerge only at larger scales from interactions of atomic building blocks. Accurate predictions at large scales are essential to move beyond trial and error approaches. Supercomputers have helped solve such problems for 50 years. Two studies from the University of Washington highlight newer techniques. One study applied artificial intelligence to simulate stacked sheets of atoms and reveal complex quantum behaviors. The other employed quantum computers to establish a loop that improves itself by discovering materials for future quantum computers. Artificial intelligence and quantum computing address different simulation challenges. Trained artificial intelligence models serve as surrogates for supercomputers, modeling large systems from limited data and enabling study of stacked atom sheets that produce novel phenomena. Quantum computers naturally handle quantum effects such as entanglement and were used to examine a Laughlin state. The researchers intend to expand datasets and integrate the tools into a hybrid system. The work was supported by Amazon and the Department of Energy.