Wyślij wiadomość: Robust reinforcement learning with Bayesian optimisation and quadrature