Machine learning one-dimensional spinless trapped fermionic systems with neural-network quantum states

Abstract

We compute the ground-state properties of fully polarized, trapped, one-dimensional fermionic systems interacting through a gaussian potential. We use an antisymmetric artificial neural network, or neural quantum state, as an ansatz for the wavefunction and use machine learning techniques to variationally minimize the energy of systems from 2 to 6 particles. We provide extensive benchmarks with other many-body methods, including exact diagonalisation and the Hartree-Fock approximation. The neural quantum state provides the best energies across a wide range of interaction strengths. We find very different ground states depending on the sign of the interaction. In the non-perturbative repulsive regime, the system asymptotically reaches crystalline order. In contrast, the strongly attractive regime shows signs of bosonization. The neural quantum state continuously learns these different phases with an almost constant number of parameters and a very modest increase in computational time with the number of particles.

References in corpus (50)

• Many-Body Physics with Ultracold Gases
• Machine learning and the physical sciences
• Solving the Quantum Many-Body Problem with Artificial Neural Networks
• Theory of ultracold Fermi gases
• Deep neural network solution of the electronic Schrödinger equation
• Ab-Initio Solution of the Many-Electron Schrödinger Equation with Deep Neural Networks
• The Equation of State of a Low-Temperature Fermi Gas with Tunable Interactions
• From Few to Many: Observing the Formation of a Fermi Sea One Atom at a Time
• Symmetries and many-body excited states with neural-network quantum states
• Fermionic neural-network states for ab-initio electronic structure
• Backflow Transformations via Neural Networks for Quantum Many-Body Wave-Functions
• NetKet 3: Machine Learning Toolbox for Many-Body Quantum Systems
• Tuning the interactions of spin-polarized fermions using quasi-one-dimensional confinement
• One-dimensional mixtures of several ultracold atoms: a review
• Solving the Bose-Hubbard model with machine learning
• Imaging the Wigner Crystal of Electrons in One Dimension
• Ab-initio quantum chemistry with neural-network wavefunctions
• Wigner Molecules in Nanostructures
• Diffusion quantum Monte Carlo study of three-dimensional Wigner crystals
• Fermionic Wave Functions from Neural-Network Constrained Hidden States
• Variational Monte Carlo calculations of $\mathbf{A}\le 4$ nuclei with an artificial neural-network correlator ansatz
• Theory of spinor Fermi and Bose gases in tight atom waveguides
• Effective interactions, Fermi-Bose duality, and ground states of ultracold atomic vapors in tight de Broglie waveguides
• Structure of fermion nodes and nodal cells
• Machine learning the deuteron
• Spatial ordering of charge and spin in quasi one-dimensional Wigner molecules
• Wave function Ansatz (but Periodic) Networks and the Homogeneous Electron Gas
• Efimov Physics and Connections to Nuclear Physics
• Method to solve quantum few-body problems with artificial neural networks
• Natural Orbitals and Occupation Numbers for Harmonium: Fermions vs. Bosons
• Bose-Fermi dualities for arbitrary one-dimensional quantum systems in the universal low energy regime
• Many-body wavefunctions for normal liquid ${}^3$He
• Strongly interacting trapped one-dimensional quantum gases: an exact solution
• An efficient sampling algorithm for Variational Monte Carlo
• Towards a nonequilibrium Green's function description of nuclear reactions: one-dimensional mean-field dynamics
• Nuclei with up to $\mathbf{A}=6$ nucleons with artificial neural network wave functions
• Quantum correlations and degeneracy of identical bosons in a 2D harmonic trap
• Universal duality transformations in interacting one-dimensional quantum systems
• Number-parity effect for confined fermions in one dimension
• Exactly solvable model of two trapped quantum particles interacting via finite-range soft-core interactions
• Scalable neural networks for the efficient learning of disordered quantum systems
• Automatic Differentiable Monte Carlo: Theory and Application
• One-dimensional few-electron effective Wigner crystal in quantum and classical regimes
• Convergence to the fixed-node limit in deep variational Monte Carlo
• Static and dynamic properties of a few spin $1/2$ interacting fermions trapped in an harmonic potential
• Orbital-dependent backflow wave functions for real-space quantum Monte Carlo
• Reduced density matrix approach to ultracold few-fermion systems in one dimension
• Solving the nuclear pairing model with neural network quantum states
• Supervised learning of few dirty bosons with variable particle number
• Variational ansatz for $p$-wave fermions confined in a one-dimensional harmonic trap

Cited by in corpus (1)

• Second-order optimisation strategies for neural network quantum states