Sandip Sen helped establish multi-agent learning as a coordination problem, studying how adaptive agents can learn complementary behavior, reciprocity, cooperation, and trust.
Michael Littman's 1994 Markov-games paper extended reinforcement learning beyond a single adaptive learner by modeling environments in which other agents also choose actions and change the learner's world.
Salesforce's CodeRL treated a code language model as an actor and learned critic models from unit-test outcomes, bringing functional execution feedback into both training and inference for program generation.