Trust Me, I Know the Way: Predictive Uncertainty in the Presence of Shortcut Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912308840628224 |
|---|---|
| author | Wimmer, Lisa Bischl, Bernd Bothmann, Ludwig |
| author_facet | Wimmer, Lisa Bischl, Bernd Bothmann, Ludwig |
| contents | The correct way to quantify predictive uncertainty in neural networks remains a topic of active discussion. In particular, it is unclear whether the state-of-the art entropy decomposition leads to a meaningful representation of model, or epistemic, uncertainty (EU) in the light of a debate that pits ignorance against disagreement perspectives. We aim to reconcile the conflicting viewpoints by arguing that both are valid but arise from different learning situations. Notably, we show that the presence of shortcuts is decisive for EU manifesting as disagreement. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_09137 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Trust Me, I Know the Way: Predictive Uncertainty in the Presence of Shortcut Learning Wimmer, Lisa Bischl, Bernd Bothmann, Ludwig Machine Learning The correct way to quantify predictive uncertainty in neural networks remains a topic of active discussion. In particular, it is unclear whether the state-of-the art entropy decomposition leads to a meaningful representation of model, or epistemic, uncertainty (EU) in the light of a debate that pits ignorance against disagreement perspectives. We aim to reconcile the conflicting viewpoints by arguing that both are valid but arise from different learning situations. Notably, we show that the presence of shortcuts is decisive for EU manifesting as disagreement. |
| title | Trust Me, I Know the Way: Predictive Uncertainty in the Presence of Shortcut Learning |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2502.09137 |