07.05.2023 - By Quanta Magazine
To improve reinforcement learning algorithms, computer scientists are training them as if they were human.