Simulating the nonequilibrium Green’s function by dynamic mode decomposition and operator learning Computing the numerical solution of the Kadanoff-Baym equations (KBEs), a set of nonlinear integral differential equations satisfied by the two-time Green’s functions derived from many-body perturbation theory for a quantum many-body system away from equilibrium, is a challenging task. In this talk, I will report our recent efforts on extrapolating the two-time Green’s function by applying dynamic mode decomposition (DMD) and recurrent neural networks (RNN)-based operator learning. These methods require constructing models from the numerical solution of the KBE within a small time window to extrapolate both the time-diagonal and off-diagonal elements of the Green’s function. We demonstrate the efficiency and accuracy of these approaches by applying it to Hubbard model problems.
