Discovering Quantum Phase Transitions with Fermionic Neural Networks
Authors:
Preprint
Abstract
Deep neural networks have been extremely successful as highly accurate wave function ans\"atze for variational Monte Carlo calculations of molecular ground states. We present an extension of one such ansatz, FermiNet, to calculations of the ground states of periodic Hamiltonians, and study the homogeneous electron gas. FermiNet calculations of the ground-state energies of small electron gas systems are in excellent agreement with previous initiator full configuration interaction quantum Monte Carlo and diffusion Monte Carlo calculations. We investigate the spin-polarized homogeneous electron gas and demonstrate that the same neural network architecture is capable of accurately representing both the delocalized Fermi liquid state and the localized Wigner crystal state. The network is given no \emph{a priori} knowledge that a phase transition exists, but converges on the translationally invariant ground state at high density and spontaneously breaks the symmetry to produce the crystalline ground state at low density.Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Deep-neural-network solution of the electronic Schrödinger equationNature Chemistry, 2020
- Restricted Boltzmann machines in quantum physicsNature Physics, 2019
- Machine learning in electronic-quantum-matter imaging experimentsNature, 2019
- Neural-network quantum state tomographyNature Physics, 2018
- Many-Body Methods in Chemistry and Physics2009
- Quantum Theory of the Electron Liquid2005
- Approximation capabilities of multilayer feedforward networksNeural Networks, 1991
- Approximation by superpositions of a sigmoidal functionMathematics of Control, Signals, and Systems, 1989
- Error estimates on averages of correlated dataThe Journal of Chemical Physics, 1989
- On the eigenfunctions of many‐particle systems in quantum mechanicsCommunications on Pure and Applied Mathematics, 1957