Research map / Multi-agent systems and games / Cooperative multi-agent RL
Correcting experience replay for multi-agent communication
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Add communication correction that relabels messages under current policies | MADDPG and MAAC · Cooperative Communication with 5 landmarks | reward | Helped |
| Add training-iteration fingerprint to the critic | MADDPG · Cooperative Communication with 5 landmarks | reward | No clear effect |
| Replace first-step correction with ordered communication correction (ablation) | MADDPG+FCC · Hierarchical Communication | reward | Helped |
| Apply first-step communication correction (ablation) | MADDPG · Hierarchical Communication | reward | No clear effect |
| Apply communication correction to the Allies | MADDPG · Covert Communication | Allies' reward | Helped |
| Apply communication correction to the Adversary | MADDPG · Covert Communication | Allies' reward | Hurt |
| Add communication correction that relabels messages under current policies | MADDPG · Cooperative Communication with Multiple Targets | reward | Helped |
| Account for message drops in the relabelling communication model | MADDPG+CC · Cooperative Communication with dropped messages | reward | Helped |
| Ignore the message-drop probability during relabelling (ablation) | MADDPG+CC · Cooperative Communication with dropped messages | reward | Hurt |
| Train with messages dropped at the stated rate | MADDPG+CC · Cooperative Communication with dropped messages | robustness to message drops | Helped |
Closest papers by meaning
- Multi-Agent Reinforcement Learning with Communication-Constrained Priors (NeurIPS 2025 · similarity 0.89)
- Sample-Efficient Multiagent Reinforcement Learning with Reset Replay (ICML 2024 · similarity 0.88)
- Efficient Communication in Multi-Agent Reinforcement Learning with Implicit Consensus Generation (AAAI 2025 · similarity 0.87)
- Robust Communicative Multi-Agent Reinforcement Learning with Active Defense (AAAI 2024 · similarity 0.87)
- Learning Multi-Agent Communication with Contrastive Learning (ICLR 2024 · similarity 0.87)
- Bridging Training and Execution via Dynamic Directed Graph-Based Communication in Cooperative Multi-Agent Systems (AAAI 2025 · similarity 0.87)
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.