Trust Me, I Know the Way: Predictive Uncertainty in the Presence of Shortcut Learning
MCML Authors
Lisa Wimmer
Dr.
* Former Member
Bernd Bischl
Prof. Dr.
Director
Ludwig Bothmann
Dr.
Abstract
Lisa Wimmer
Dr.
* Former Member
Bernd Bischl
Prof. Dr.
Director
Ludwig Bothmann
Dr.
Abstract
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.
inproceedings WBB25
SCSL @ICLR 2025
Workshop on Spurious Correlation and Shortcut Learning: Foundations and Solutions at the 13th International Conference on Learning Representations. Singapore, Apr 24-28, 2025.Authors
L. Wimmer • B. Bischl • L. BothmannLinks
URLResearch Area
BibTeXKey: WBB25