Learning policy representations in multiagent systems

Read on openai.com