Ab initio quantum chemistry with neural-network wavefunctions

Deep learning methods outperform human capabilities in pattern recognition and data processing problems and now have an increasingly important role in scientific discovery. A key application of machine learning in molecular science is to learn potential energy surfaces or force fields from ab initio solutions of the electronic Schrodinger equation using data sets obtained with density functional theory, coupled cluster or other quantum chemistry (QC) methods. In this Review, we discuss a complementary approach using machine learning to aid the direct solution of QC problems from first principles. Specifically, we focus on quantum Monte Carlo methods that use neural-network ansatzes to solve the electronic Schrodinger equation, in first and second quantization, computing ground and excited states and generalizing over multiple nuclear configurations. Although still at their infancy, these methods can already generate virtually exact solutions of the electronic Schrodinger equation for small systems and rival advanced conventional QC methods for systems with up to a few dozen electrons.

Automated summary

Deep learning methods have surpassed human capabilities in pattern recognition and data processing, and have become increasingly important in scientific discovery. In molecular science, a key application of machine learning is to learn potential energy surfaces or force fields from ab initio solutions of the electronic Schrodinger equation obtained with various quantum chemistry methods. This review discusses a complementary approach that uses machine learning to directly solve quantum chemistry problems from first principles, focusing on quantum Monte Carlo methods with neural-network ansatzes to solve the electronic Schrodinger equation.