Continuous-state and discrete-action control on Acrobot v1
495.4Final Average RewardIndep. QHD
Evaluation Results
| Method | Links | |
|---|---|---|
| Indep. QHDEncoder Setting=Hetero2026.05 | 495.4 | |
| Indep. QHDEncoder Setting=Homo2026.05 | 306.1 | |
| Trunc. FedAvg-QHDEncoder Setting=Hetero2026.05 | 280.8 | |
| Trunc. FedAvg-QHDEncoder Setting=Homo2026.05 | 106.5 | |
| FedQHDEncoder Setting=Homo2026.05 | 105 | |
| FedQHDEncoder Setting=Hetero2026.05 | 102.5 | |
| Oracle QHD†Encoder Setting=Hetero2026.05 | 101.2 | |
| Oracle QHD†Encoder Setting=Homo2026.05 | 93.6 | |
| Distill. FedDQNEncoder Setting=Hetero2026.05 | 89.2 | |
| Distill. FedDQNEncoder Setting=Homo2026.05 | 87.5 | |
| FedAvg-DQNEncoder Setting=Homo2026.05 | 85.3 | |
| Oracle DQN†Encoder Setting=Hetero2026.05 | 79.8 | |
| Oracle DQN†Encoder Setting=Homo2026.05 | 75.3 |