Combining trajectory sampling and deep Gaussian covariance network (DGCN) enhances data-efficient policy search in model-based reinforcement learning.
