Vision-Language Navigation Failure Identification on ApexNav
20.4Perception Failure RateRandom
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| RandomAugmentation=+OA2026.04 | 20.4 | 16.4 | 19.2 | 24.8 | |
| BehAVExplorAugmentation=+OA2026.04 | 40.9 | 46.4 | 35.6 | 43.2 | |
| VLATestAugmentation=+OA2026.04 | 51.8 | 54.1 | 45.1 | 52.1 | |
| CanTest2026.04 | 72.2 | 66.3 | 52.5 | 59.5 |