English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Many important computer vision applications are naturally formulated asregression problems. Within medical imaging, accurate regression models havethe potential to automate various tasks, helping to lower costs and improvepatient outcomes. Such safety-critical deployment does however require reliableestimation of model uncertainty, also under the wide variety of distributionshifts that might be encountered in practice. Motivated by this, we set out toinvestigate the reliability of regression uncertainty estimation methods undervarious real-world distribution shifts. To that end, we propose an extensivebenchmark of 8 image-based regression datasets with different types ofchallenging distribution shifts. We then employ our benchmark to evaluate manyof the most common uncertainty estimation methods, as well as twostate-of-the-art uncertainty scores from the task of out-of-distributiondetection. We find that while methods are well calibrated when there is nodistribution shift, they all become highly overconfident on many of thebenchmark datasets. This uncovers important limitations of current uncertaintyestimation methods, and the proposed benchmark therefore serves as a challengeto the research community. We hope that our benchmark will spur more work onhow to develop truly reliable regression uncertainty estimation methods. Codeis available at https://github.com/fregu856/regression_uncertainty.

How Reliable is Your Regression Model's Uncertainty Under Real-World Distribution Shifts?