In a pair of papers accepted to the International Conference on Learning Representations (ICLR) 2020, MIT researchers investigated new ways to motivate software agents to…