Visual Planning for Assistance
50.6Mean Accuracy (mAcc)VLaMP
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| VLaMPPlanning Horizon (T)=1, Training Scheme=Fully Supervised2026.05 | 50.6 | 10.3 | — | |
| VidAssist w/ RECIPE-RLPlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 39.9 | 9.3 | — | |
| VidAssistPlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 38.7 | 8.7 | — | |
| VidAssist w/ BasePlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 36.5 | 2 | — | |
| VLaMPPlanning Horizon (T)=3, Training Scheme=Fully Supervised2026.05 | 35.3 | — | 44 | |
| DDNPlanning Horizon (T)=1, Training Scheme=Fully Supervised2026.05 | 33.4 | 6.8 | — | |
| VLaMPPlanning Horizon (T)=4, Training Scheme=Fully Supervised2026.05 | 31.7 | 4.4 | 43.4 | |
| VidAssist w/ RECIPE-RLPlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 29.9 | — | 35.7 | |
| LLM Agent BaselinePlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 28.7 | 3.1 | — | |
| VidAssistPlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 28.5 | — | 44.1 | |
| VidAssist w/ BasePlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 27.3 | — | 33 | |
| LTAPlanning Horizon (T)=1, Training Scheme=Fully Supervised2026.05 | 26.9 | 2.4 | — | |
| VidAssist w/ RECIPE-RLPlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 26 | 4.8 | 34.5 | |
| DDNPlanning Horizon (T)=3, Training Scheme=Fully Supervised2026.05 | 25.8 | — | 35.2 | |
| LLM BaselinePlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 25.8 | 2.1 | — | |
| VidAssistPlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 25.8 | 4.6 | 46.8 | |
| LLM Agent BaselinePlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 25.6 | — | 29.9 | |
| VidAssist w/ BasePlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 24.3 | 1 | 32 | |
| DDNPlanning Horizon (T)=4, Training Scheme=Fully Supervised2026.05 | 24.1 | 3.6 | 37 | |
| LTAPlanning Horizon (T)=3, Training Scheme=Fully Supervised2026.05 | 24 | — | 35.2 | |
| LLM BaselinePlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 23.2 | — | 27.7 | |
| LTAPlanning Horizon (T)=4, Training Scheme=Fully Supervised2026.05 | 21.7 | 1.2 | 36.8 | |
| LLM Agent BaselinePlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 20 | 0.8 | 35.6 | |
| LLM BaselinePlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 18.9 | 0.4 | 33 | |
| Random w/ goalPlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 13.4 | — | 23.6 | |
| Random w/ goalPlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 13.2 | 0.3 | — | |
| Random w/ goalPlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 12.7 | 0 | 27.8 | |
| Most Probable w/ goalPlanning Horizon (T)=1, Training Scheme=Fully Supervised2026.05 | 12.4 | 2.4 | — | |
| Most ProbablePlanning Horizon (T)=1, Training Scheme=Fully Supervised2026.05 | 10.4 | 1.7 | — | |
| Most Probable w/ goalPlanning Horizon (T)=3, Training Scheme=Fully Supervised2026.05 | 8.9 | — | 15.5 | |
| Most Probable w/ goalPlanning Horizon (T)=4, Training Scheme=Fully Supervised2026.05 | 7.9 | 1.5 | 20.5 | |
| Most ProbablePlanning Horizon (T)=3, Training Scheme=Fully Supervised2026.05 | 6.1 | — | 9.9 | |
| Most ProbablePlanning Horizon (T)=4, Training Scheme=Fully Supervised2026.05 | 5.5 | 1.3 | 13.9 | |
| RandomPlanning Horizon (T)=1, Training Scheme=Zero-Shot2026.05 | 0.9 | 0 | — | |
| RandomPlanning Horizon (T)=3, Training Scheme=Zero-Shot2026.05 | 0.9 | — | 1.5 | |
| RandomPlanning Horizon (T)=4, Training Scheme=Zero-Shot2026.05 | 0.9 | 0 | 1.9 |