Discovering Quantum Phase Transitions with Fermionic Neural Networks
This work addresses the challenge of simulating quantum phase transitions in periodic systems for condensed matter physics, representing an incremental advancement in applying neural networks to quantum many-body problems.
The authors extended the FermiNet neural network to model periodic Hamiltonians, specifically the homogeneous electron gas, achieving ground-state energy results that closely match previous high-accuracy quantum Monte Carlo methods. They demonstrated that the same architecture can accurately represent both Fermi liquid and Wigner crystal states, spontaneously capturing the phase transition without prior knowledge.
Deep neural networks have been extremely successful as highly accurate wave function ansätze 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.