Evaluating analytic gradients of pulse programs on quantum computers

Abstract: Parametrized pulse programs running on quantum hardware can be differentiated via the stochastic parameter-shift (SPS) rule. We overcome the intrinsically approximate nature of SPS by introducing a new analytic method for computing gradients of pulse programs, that we coin ODEgen. In this new method, we find effective generators of pulse gates using a differentiable ordinary differential equation (ODE) solver. These effective generators inform parameter-shift rules that can be evaluated on quantum hardware. We showcase simulated VQE examples with realistic superconducting transmon systems, for which we obtain lower energies with fewer quantum resources using ODEgen over SPS. We further demonstrate a pulse VQE run with gradients computed via ODEgen entirely on quantum hardware.