Automated Incentive Token Exchange for Cooperative Multi-Agent Reinforcement Learning
MEDIATE proposes an automated mechanism to derive dynamic agent-based incentivization tokens and a consensus protocol to ensure global convergence of the tokens, enabling emergent cooperation between self-interested agents in various social dilemma environments.