Failure Detection on Close Box Combined
100TPRTemporal Non-Distr. Min.
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Temporal Non-Distr. Min.Failure Detector Category=Diffusion2026.03 | 100 | 77 | 85 | |
| DDPM Loss (Eq. A.1)Failure Detector Category=Diffusion2026.03 | 100 | 91 | 94 | |
| Temporal DDPM LossFailure Detector Category=Diffusion2026.03 | 100 | 79 | 86 | |
| CLIP PretrainedFailure Detector Category=Embed.2026.03 | 100 | 68 | 79 | |
| ResNet PretrainedFailure Detector Category=Embed.2026.03 | 100 | 68 | 79 | |
| GPT-4o Image QAFailure Detector Category=VLM2026.03 | 100 | 0 | 29 | |
| Gemini 1.5 Pro Image QAFailure Detector Category=VLM2026.03 | 100 | 0 | 29 | |
| Gemini 1.5 Pro Video QAFailure Detector Category=VLM2026.03 | 100 | 55 | 68 | |
| SentinelFailure Detector Category=Combined, Configuration=STAC MMD* + GPT-4o Video QA*2026.03 | 100 | 87 | 91 | |
| STAC For. KLFailure Detector Category=STAC2026.03 | 99 | 89 | 92 | |
| STAC MMD*Failure Detector Category=STAC2026.03 | 99 | 94 | 96 | |
| Policy EncoderFailure Detector Category=Embed.2026.03 | 94 | 70 | 78 | |
| STAC Rev. KLFailure Detector Category=STAC2026.03 | 93 | 96 | 95 | |
| Temporal Diffusion Recon.Failure Detector Category=Diffusion2026.03 | 92 | 97 | 95 | |
| GPT-4o Video QA*Failure Detector Category=VLM2026.03 | 77 | 91 | 87 | |
| Claude 3.5 Sonnet Image QAFailure Detector Category=VLM2026.03 | 77 | 7 | 27 | |
| Claude 3.5 Sonnet Video QAFailure Detector Category=VLM2026.03 | 55 | 35 | 40 | |
| Diffusion Recon.Failure Detector Category=Diffusion2026.03 | 37 | 96 | 76 | |
| Diffusion Output VarianceFailure Detector Category=Diffusion2026.03 | 26 | 96 | 72 |