TY - RPRT TI - A Stochastic Derivative-Free Optimization Method with Importance Sampling: Theory and Learning to Control AU - Adel Bibi AU - El Houcine Bergou AU - Ozan Sener AU - Bernard Ghanem AU - Peter Richtárik PY - 2020 UR - https://arxiv.org/abs/1902.01272 ID - 1902.01272 ER -