AI and the World's Hard Problems · Quantum physics & quantum computing · arXiv · 2026-10-21
Neural Network Fractional Quantum Hall States with Learnable Orbital Pair Matrix Factors: Balancing Physical Priors and Flexibility
Neural network fractional quantum Hall states promise to capture Landau level mixing, but incorporating physical priors has been hard to reconcile with network flexibility. The proposed orbital pair matrix factor (MHPO-M) ansatz replaces the fixed Laughlin structure with learnable pair factors that depend on both electron configurations and orbitals. Across ν=1/5, 1/3, and 2/5 it achieves the lowest variational energy among the tested variants, and a factorized orbital readout enables scaling to 35 electrons with energies competitive with trends inferred from three-LL exact diagonalization.