For the past decade, quantum materials, where complex phenomena emerge from complex orbital, charge, lattice and spin interactions, have been a source of enormous fundamental breakthroughs. Quantum materials are difficult if not impossible to understand solely using existing simulation and analytical techniques, therefore, insights from experimental data are of critical importance. Machine learning continues to advance as a powerful tool for understanding and designing materials. These methods have been highly successful in improving atomistic simulations, materials design and discovery, literature information extraction, and quantum information systems. However, the following challenges remain unresolved when merging machine learning methods with experimental data: 1) Experimental data can have noise from various sources. 2) In comparison to high-throughput computational results, experimental data is often small and scattered. 3) Experimental data often disagree with simulation or computational results.

Given the vast opportunities that machine learning may bring experimental investigations of quantum materials, we feel it obligatory and timely to organize a symposium to address the following challenges and bottlenecks of applying state-of-the-art machine learning architectures to address key open questions in quantum materials: How do we apply machine learning to noisy, scarce experimental data, especially when there is disagreement with computational data? How do we extract key insights on quantum materials from experimental data that cannot be accessed from conventional manual data analysis?

This symposium will highlight recent progress in applying machine learning to various types of materials characterization techniques: neutron and x-ray scattering, optical spectroscopies, angular-resolved photoemission spectroscopy (ARPES), free electron laser, among other emerging novel spectroscopies. We will cover diverse quantum materials, such as novel topological insulators, semimetals and metals, in bulk, thin film, 2D and 1D form. We will emphasize recent progress in machine learning techniques relevant to noise reduction and inferring missing or the corruption of data. We will focus on applications of machine learning to augment experimental data for novel quantum materials, as well as “hot off the press” characterization and analysis tools for quantum materials. This symposium will provide an interactive, widely-accessible forum for materials scientists to get up to speed on the exciting recent progress of machine learning and quantum materials. To ensure cross-fertilization of these new techniques and approaches, sessions will be organized by scientific theme rather than material category. Additional sessions will focus on recent methodological advances of the machine learning capabilities to probe the charge, spin or lattice degrees of freedom.

electron energy loss spectroscopy (EELS) machine learning neutron scattering quantum materials Raman spectroscopy spectroscopy transmission electron microscopy (TEM) x-ray photoelectron spectroscopy (XPS)