Robotic Manipulation on Simulation open-drawer
0.001Success Rate P-ValueReward Machine vs LLM-Guided Subgoal agent
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Reward Machine vs LLM-Guided Subgoal agentEvaluation Protocol=Wilcoxon signed-rank test, Agent Architecture=PPO, Max Episode Length=1002026.03 | 0.001 | 0.001 | |
| LEAGUE sparse agent vs LLM-Guided Subgoal agentEvaluation Protocol=Wilcoxon signed-rank test, Agent Architecture=PPO, Max Episode Length=1002026.03 | 0.002 | 0.002 | |
| LEAGUE sparse agent vs Reward MachineEvaluation Protocol=Wilcoxon signed-rank test, Agent Architecture=PPO, Max Episode Length=1002026.03 | 1 | 0.5 |