Multi-Agent Reinforcement Learning on Battle N=10,000 (Expert)
88.9Normalized ReturnMF-Diffuser
Evaluation Results
| Method | Links | |
|---|---|---|
| MF-Diffuser2026.05 | 88.9 | |
| MF-CDMs-RL2026.05 | 80.2 | |
| Oryx2026.05 | 78.5 | |
| MFQ-Offline2026.05 | 73.5 | |
| DoF2026.05 | 64.8 | |
| Indep. Diffuser2026.05 | 59.3 | |
| MA-TD3+BC2026.05 | 35.2 | |
| OMAR2026.05 | 21.8 | |
| MADiffattention=chunked-attention (max 12-agent windows)2026.05 | 9.8 | |
| Joint Diffuserattention=chunked-attention (block size 32)2026.05 | 5.8 |