Evidence-Guided Unknown Rejection for High-Confidence Near-Known Unknowns
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911692642844672 |
|---|---|
| author | Chen, Xi Xiao, Yingjun Fang, Gang |
| author_facet | Chen, Xi Xiao, Yingjun Fang, Gang |
| contents | Open-set recognition systems face a neglected failure mode: high-confidence near-known unknowns, which lie outside the known label set but are close enough to known classes that a closed-set classifier accepts them with high confidence. We show that this failure is widespread across scalar-threshold methods, including recent post-hoc detectors, and that stronger encoders can amplify rather than remove the risk. We propose EGUR-A, which changes the decision from ``is this sample's score high enough?'' to ``does this predicted known class have sufficient evidence to accept this sample?'' EGUR-A combines class-conditional local acceptance evidence with global residual evidence, and selects their relative weight from known-sample statistics without unknown validation data. Across CUB, FGVC-Aircraft, and ImageNet-hard, EGUR-A substantially reduces high-confidence false known acceptance at matched known-rejection operating points. The result is not a stronger threshold; it is a different question: whether a known class is entitled to accept a sample. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_17818 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Evidence-Guided Unknown Rejection for High-Confidence Near-Known Unknowns Chen, Xi Xiao, Yingjun Fang, Gang Computer Vision and Pattern Recognition Open-set recognition systems face a neglected failure mode: high-confidence near-known unknowns, which lie outside the known label set but are close enough to known classes that a closed-set classifier accepts them with high confidence. We show that this failure is widespread across scalar-threshold methods, including recent post-hoc detectors, and that stronger encoders can amplify rather than remove the risk. We propose EGUR-A, which changes the decision from ``is this sample's score high enough?'' to ``does this predicted known class have sufficient evidence to accept this sample?'' EGUR-A combines class-conditional local acceptance evidence with global residual evidence, and selects their relative weight from known-sample statistics without unknown validation data. Across CUB, FGVC-Aircraft, and ImageNet-hard, EGUR-A substantially reduces high-confidence false known acceptance at matched known-rejection operating points. The result is not a stronger threshold; it is a different question: whether a known class is entitled to accept a sample. |
| title | Evidence-Guided Unknown Rejection for High-Confidence Near-Known Unknowns |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.17818 |