Game-Theoretic Multi-Agent Reinforcement Learning