| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Action anticipation | EPIC (val) | Top-5 Action Accuracy40.2 | 28 | |
| Generalization to Novel Objects | EPIC (test) | KLD1.197 | 8 | |
| Affordance Grounding | EPIC (test) | KLD1.209 | 8 | |
| Parallel Execution Prediction | EPIC 2-Body Problem (test) | Coverage100 | 8 | |
| Video Recognition | EPIC | Accuracy63.3 | 8 | |
| Open-vocabulary object recognition | EPIC100-OV | Top-1 Accuracy (Base)52.2 | 8 | |
| Generalization to novel objects | EPIC novel objects | KLD1.249 | 8 | |
| Grounded affordance prediction | EPIC (seen classes) | KLD1.258 | 8 | |
| Action Recognition | EPIC100-OV (HM) | Verb Top-1 Acc50.3 | 3 | |
| Action Recognition | EPIC100-OV (Novel) | Verb Top-1 Acc0.414 | 3 | |
| Action Recognition | EPIC100 OV Closed | Verb Top-1 Acc64.1 | 3 | |
| Autonomous Exploration | EPIC Garage | Time (s)657.85 | 2 | |
| Parallel Execution Prediction | EPIC 3-Body Problem (test) | Coverage0.83 | 1 |