Return to Article Details Multi-agent training for Pommerman: Curriculum learning and population-based self-play approach Download Download PDF