Acceleration of Quantum Materials Discovery Driven by Machine Learning and Text Data Mining

Abstract

Traditional materials discovery methods involve massive expense of calculation or experiments due to the complexity of experimental condition combinations. Machine learning(ML) has recently demonstrated powerful capabilities for finding useful hidden information from huge and noisy datasets. Currently there are not sufficient experimental records to harness the power of machine learning to speed up materials discovery. Thus we are building an infrastructure for data mining from the literature, using Natural Language Processing (NLP).