Cooperative Multi-agent Reinforcement Learning (coo-MARL) on QMIX Poisoned
98.6Average Winning RateSHINE
Evaluation results
| Method | Links | |
|---|---|---|
| SHINE2026.05 | 98.6 | |
| BehaviorGuard2026.05 | 96.9 | |
| BIRD2026.05 | 95.1 | |
| STRIP2026.05 | 85.3 | |
| Direct retraining2026.05 | 82.6 | |
| FeatureRE2026.05 | 33.2 | |
| NC2026.05 | 25.9 | |
| Original2026.05 | 21.9 |