Classic Discrete Control on MountainCar v0
200Mean Episodic ReturnIndep. QHD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Indep. QHDEncoder Setting=Homo2026.05 | 200 | — | |
| Indep. QHDEncoder Setting=Hetero2026.05 | 200 | — | |
| Distill. FedDQNEncoder Setting=Hetero2026.05 | 174.3 | — | |
| Trunc. FedAvg-QHDEncoder Setting=Hetero2026.05 | 172.9 | — | |
| FedAvg-DQNEncoder Setting=Homo2026.05 | 167.8 | — | |
| FedQHDEncoder Setting=Hetero2026.05 | 162.3 | — | |
| Distill. FedDQNEncoder Setting=Homo2026.05 | 156 | — | |
| Trunc. FedAvg-QHDEncoder Setting=Homo2026.05 | 148.3 | — | |
| FedQHDEncoder Setting=Homo2026.05 | 143.7 | — | |
| Oracle QHD†Encoder Setting=Homo2026.05 | 141.3 | — | |
| Oracle QHD†Encoder Setting=Hetero2026.05 | 140 | — | |
| MLP-PPOArchitecture=MLP2026.04 | 128 | — | |
| Oracle DQN†Encoder Setting=Homo2026.05 | 125.7 | — | |
| Oracle DQN†Encoder Setting=Hetero2026.05 | 122.8 | — | |
| INT4-PPOPrecision=INT42026.04 | 122 | — | |
| INT8-PPOPrecision=INT82026.04 | 118 | — | |
| FP16-PPOPrecision=FP162026.04 | 115 | — | |
| BitRL-1bitPrecision=1-bit2026.04 | 108 | 106.1 |