Efficient VQE with the perturbation method

Quantum-classical hybrid approaches that harness computational power of early NISQ devices have recently been demonstrated on multiple platforms and VQE is perhaps one of the most widely known. In a typical VQE run, one approximates the ground state energy of a target molecular system by iteratively creating an ansatz ground state and evaluating its energy on a quantum computer. To boost the quality of the energy estimate and obtain faster convergence, it is crucial that one expends as little quantum resource as possible, since each quantum operation incurs computational error. In this talk, we present a perturbation-based method, namely, hybrid second-order Møllar-Plesset perturbation (HMP2), capable of guiding the development of a good ansatz state while significantly reducing the quantum resources required to estimate the energy.