Control as Probabilistic Inference as an Emergent Communication Mechanism in Multi-Agent Reinforcement Learning

Open in new window