Researchers develop technique to increase sample efficiency in reinforcement learning

Google, University of Oxford, and UC Berkeley researchers propose in a preprint paper a sample-efficient reinforcement learning technique.