[NIPS 2019] Multi-Agent Common Knowledge Reinforcement Learning笔记

文章目录

• 前言
• Introduction
• Problem setting
• • Common knowledge
  • Learning under common knowledge (LuCK)
  • Field-of-view common knowledge
• Multi-Agent Common Knowledge Reinforcement Learning
• • Pairwise MACKRL
  • Training
• 实验
• • Single-step matrix game
  • StarCraft II micromanagement
• Conclusion and Future Work