Machine learning (ML) has demonstrated great power in materials science in the past few years. In this seminar, I will introduce how ML can augment the x-ray, neutron, and electron scattering analysis by revealing the buried information and accelerating materials design. After a brief overview of the machine learning and its general applications to spectroscopies, I'll provide three examples in elastic scattering, inelastic scattering, and time-resolved spectroscopy, referring to an increased level of difficulty carrying out machine learning with plenty of data, limited data, and no data. In elastic scattering, we introduce how ML can lead to robust structural information extractions in heterostructures beyond analytical fitting models; for inelastic scattering, we introduce an efficient predictor of vibrational density-of-states in solids even in alloy space with ab initio accuracy but low computational cost by using crystal symmetry to augment data volume . Finally, we further show how ML can assist ultrafast diffraction analysis that leads to a panoramic mapping of thermal transport. We conclude by envisioning a variety of problems machine learning may solve in neutron, x-ray and electron spectroscopic researches .
 Direct Prediction of Phonon Density of States With Euclidean Neural Networks, Adv. Sci. 8, 2004214 (2021)
 Machine learning on neutron and x-ray scattering and spectroscopies, Chem. Phys. Rev. 2, 031301 (2021)
Mingda Li is the Norman C. Rasmussen Assistant Professor at MIT Nuclear Science and Engineering Department. He completed his B.S in Engineering Physics from Tsinghua University in 2009, PhD in Nuclear Science and Engineering from MIT in 2015, and did his postdoc at MIT Mechanical Engineering. His research focuses on designing new paradigms to better measure materials, with a focus on acquiring microscopic interaction information of quantum materials that were not readily measurable with current techniques.
Zoom link for invitation talk at Noon-1pm, September 17, 2021:
Zoom link for one-on-one meeting on September 17, 2021 (individual meeting between UMass faculty and Mingda)