Research Director

Naoto Tsuji

Center for Emergent Matter Science (CEMS), RIKEN
Research Scientist

Outline

In this project, we aim to establish a general framework of a new mathematical approach to optimize the efficiency of quantum many-body calculations by using machine learning methods. More precisely, we construct a machine learning model to infer an optimal choice of wave function bases for a quantum Monte Carlo algorithm used in the dynamical mean-field theory. We apply this approach to enable a high-accuracy calculation for multi-orbital and/or large-cluster material models which cannot be treated with the existing algorithms.