aiqbench analysis
This research introduces a nonparametric re-calibration algorithm using conditional kernel mean embeddings to improve uncertainty estimates in probabilistic regression without restrictive parametric assumptions. By addressing the tendency of models to produce overconfident predictions, this method offers more trustworthy interval estimates for safety-critical applications while maintaining efficient inference times. It represents a significant advancement in aligning predictive distributions with true empirical uncertainty across diverse model classes and benchmarks.
Nonparametric Distribution Regression Re-calibration
arXiv:2602.13362v2 Announce Type: replace-cross Abstract: A key challenge in probabilistic regression is ensuring that predictive distributions accurately reflect true empirical uncertainty. Minimizing overall prediction error often encourages models to prioritize informativeness over calibration, producing narrow but overconfident predictions. However, in safety-critical settings, trustworthy uncertainty estimates are often more valuable than narrow intervals. Realizing the problem, several recent works have focused on post-hoc corrections; however, existing methods either rely on weak notions of calibration (such as PIT uniformity) or impose restrictive parametric assumptions on the nature of the error. To address these limitations, we propose a novel nonparametric re-calibration algorithm based on conditional kernel mean embeddings, capable of correcting calibration error without restrictive modeling assumptions. For efficient inference with real-valued targets, we introduce a novel characteristic kernel over distributions that can be evaluated in $\mathcal{O}(n \log n)$ time for empirical distributions of size $n$. We demonstrate that our method consistently outperforms prior re-calibration approaches across a diverse set of regression benchmarks and model classes.
Abstract:A key challenge in probabilistic regression is ensuring that predictive distributions accurately reflect true empirical uncertainty. Minimizing overall prediction error often encourages models to prioritize informativeness over calibration, producing narrow but overconfident predictions. However, in safety-critical settings, trustworthy uncertainty estimates are often more valuable than narrow intervals. Realizing the problem, several recent works have focused on post-hoc corrections; however, existing methods either rely on weak notions of calibration (such as PIT uniformity) or impose restrictive parametric assumptions on the nature of the error. To address these limitations, we propose a novel nonparametric re-calibration algorithm based on conditional kernel mean embeddings, capable of correcting calibration error without restrictive modeling assumptions. For efficient inference with real-valued targets, we introduce a novel characteristic kernel over distributions that can be evaluated in $\mathcal{O}(n \log n)$ time for empirical distributions of size $n$. We demonstrate that our method consistently outperforms prior re-calibration approaches across a diverse set of regression benchmarks and model classes.
Machine Learning (stat.ML); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2602.13362 [stat.ML]
(or arXiv:2602.13362v2 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2602.13362 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 256 words
Related coverage
Fast Nonparametric Conditional Independence Testing via Two-Stage Regression
arXiv · Jun 17, 2026
Nonparametric Goodness-of-fit Testing under Covariate Shift
arXiv · Aug 6, 2026
Trustworthy Predictive Distributions for Tail Events with Semiparametric Diagnostic Transport Maps
arXiv · Sep 22, 2026
Differentially Private Nonparametric Modal Learning with Applications to Regression and Clustering
arXiv · Aug 3, 2026