Deep Gaussian Covariance Network with Trajectory Sampling for Data-Efficient Policy Search

arXiv:2403.15908v1 Announce Type: new
Abstract: Probabilistic world models increase data efficiency of model-based reinforcement learning (MBRL) by guiding the policy with their epistemic uncertainty to improve exploration and acquire new samples. Moreover, the uncertainty-aware learning procedures in probabilistic approaches lead to robust policies that are less sensitive to noisy observations compared to uncertainty unaware solutions. We propose to combine trajectory sampling and deep Gaussian covariance network (DGCN) for a data-efficient solution to MBRL problems in an optimal control setting. We …

abstract arxiv covariance cs.lg data efficiency exploration network policies policy reinforcement reinforcement learning robust samples sampling search stat.ml trajectory type uncertainty world world models