Research map / Multi-agent systems and games / Cooperative multi-agent RL
Multi-Agent Common Knowledge Reinforcement Learning
Christian Schroeder de Witt, Jakob Foerster, Gregory Farquhar, Philip Torr, Wendelin Boehmer, Shimon Whiteson
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use pairwise common-knowledge coordination with a hierarchical policy | MACKRL · 2s3z | win rate | Improved |
| Use pairwise common-knowledge coordination with a hierarchical policy | MACKRL · 3m | win rate | Improved |
| Use pairwise common-knowledge coordination with a hierarchical policy | MACKRL · 8m | win rate | Improved |
| Exploit common knowledge through pair controllers rather than independent-only policies | MACKRL · 2s3z, 3m, and 8m | sample efficiency and limit performance | Improved |
| Use MACKRL instead of Independent Actor Critic or CK-JAL | Two-agent matrix game with partial observability · Two-agent matrix game | performance | Improved |
| Use MACKRL instead of centrally executed Joint Action Learning | Two-agent matrix game · Two-agent matrix game | performance | Hurt |
| Add probabilistic common-knowledge coordination under sensor noise | MACKRL · Noisy two-agent matrix game | performance | Hurt |
| Restrict the pair selector to a fixed random subset of pairwise partitions (ablation) | MACKRL · StarCraft II SMAC tasks | sample efficiency | Hurt |
Closest papers by meaning
- Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments (NeurIPS 2017 · similarity 0.90)
- Communication-Efficient Actor-Critic Methods for Homogeneous Markov Games (ICLR 2022 · similarity 0.90)
- Automatic Grouping for Efficient Cooperative Multi-Agent Reinforcement Learning (NeurIPS 2023 · similarity 0.90)
- Integrating Suboptimal Human Knowledge with Hierarchical Reinforcement Learning for Large-Scale Multiagent Systems (NeurIPS 2024 · similarity 0.89)
- F2A2: Flexible Fully-decentralized Approximate Actor-critic for Cooperative Multi-agent Reinforcement Learning (JMLR 2023 · similarity 0.89)
- Cautiously-Optimistic Knowledge Sharing for Cooperative Multi-Agent Reinforcement Learning (AAAI 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
