Representing Quantum Many-Body States by Machine Learning

Abstract

Machine learning is used to extract essential pattern from big data. This technique can be used to extract the essential feature of quantum many-body wave function (=a vector with exponentially large dimensions), and to obtain compact representation of many-body states. In this article, we review representations of many-body states using Boltzmann machine, a type of artificial neural network. We introduce an efficient representation using restricted Boltzmann machines (RBM) and also discuss the efforts to improve the RBM wave functions.