Neural networks as low-cost surrogates for impurity solvers in quantum embedding methods
Researchers developed a neural network to replace expensive impurity solvers in dynamical mean-field theory simulations of correlated electron models. The network was trained on synthetic data generated by a more accurate solver, achieving accuracy comparable to continuous-time quantum Monte Carlo methods. This low-cost surrogate significantly speeds up DMFT calculations, reducing computation time by factors of 3.4x to 5.7x.