CarGoal1 on Safety Gymnasium
79RewardCPQ
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CPQ2026.04 | 79 | 1.42 | |
| CPQcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 79 | 1.42 | |
| DT + SASBackbone=DT, Alignment strategy (SAS)=Lyapunov-Guided2026.04 | 67 | 0.85 | |
| DT + SAScost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 67 | 0.85 | |
| CDTBackbone=CDT2026.04 | 66 | 1.21 | |
| CDTcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 66 | 1.21 | |
| CDT + SASBackbone=CDT, Alignment strategy (SAS)=Lyapunov-Guided2026.04 | 65 | 0.9 | |
| CDT + SAScost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 65 | 0.9 | |
| BEAR-Lag2026.04 | 61 | 1.13 | |
| BEAR-Lagcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 61 | 1.13 | |
| BCQ-Lag2026.04 | 47 | 0.78 | |
| BCQ-Lagcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 47 | 0.78 | |
| BC-All2026.04 | 39 | 0.33 | |
| BC-Allcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 39 | 0.33 | |
| COptiDICE2026.04 | 35 | 0.54 | |
| DCRL2026.04 | 35 | 0.88 | |
| COptiDICEcost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 35 | 0.54 | |
| BC-Safe2026.04 | 24 | 0.28 | |
| BC-Safecost thresholds=3, evaluation episodes=20, random seeds=32026.04 | 24 | 0.28 |