Researchers from Stanford University and Google DeepMind have unveiled Step-Wise Reinforcement Learning (SWiRL), a technique ...