Improving neural network performance for solving quantum sign structure

Neural quantum states have emerged as a widely used approach to the numerical study of the ground states of nonstoquastic Hamiltonians. However, existing approaches often rely on a priori knowledge of the sign structure or require a separately pretrained phase network. We introduce a modified stochastic reconfiguration method that effectively uses differing imaginary time steps to evolve the amplitude and phase. Using a larger time step for phase optimization, this method enables a simultaneous and efficient training of phase and amplitude neural networks. The efficacy of our method is demonstrated on the Heisenberg ${J}_{1}\ensuremath{-}{J}_{2}$ model.