EUVIMEDEUVIMED entdecken
Crossref

Easy Uncertainty Quantification (EasyUQ): Generating Predictive Distributions from Single-Valued Model Output

Eva-Maria Walz, Alexander Henzi, Johanna Ziegel, Tilmann Gneiting

SIAM Review · 2024-02 · 66 (1) · 91-122

DOI: 10.1137/22m1541915 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SIAM Review, online erschienen am 08.02.2024. Verfasst von Eva-Maria Walz und 3 weiteren Autor:innen. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 76 Literaturangaben. Gefördert von: Swiss National Science Foundation, Deutsche Forschungsgemeinschaft, Klaus Tschira Stiftung.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

How can we quantify uncertainty if our favorite computational tool---be it a numerical, statistical, or machine learning approach, or just any computer model---provides single-valued output only? In this article, we introduce the Easy Uncertainty Quantification (EasyUQ) technique, which transforms real-valued model output into calibrated statistical distributions, based solely on training data of model output--outcome pairs, without any need to access model input. In its basic form, EasyUQ is a special case of the recently introduced isotonic distributional regression (IDR) technique that leverages the pool-adjacent-violators algorithm for nonparametric isotonic regression. EasyUQ yields discrete predictive distributions that are calibrated and optimal in finite samples, subject to stochastic monotonicity. The workflow is fully automated, without any need for tuning. The Smooth EasyUQ approach supplements IDR with kernel smoothing, to yield continuous predictive distributions that preserve key properties of the basic form, including both stochastic monotonicity with respect to the original model output and asymptotic consistency. For the selection of kernel parameters, we introduce multiple one-fit grid search, a computationally much less demanding approximation to leave-one-out cross-validation. We use simulation examples and forecast data from weather prediction to illustrate the techniques. In a study of benchmark problems from machine learning, we show how EasyUQ and Smooth EasyUQ can be integrated into the workflow of neural network learning and hyperparameter tuning, and we find EasyUQ to be competitive with conformal prediction as well as more elaborate input-based approaches.

Förderung

  • Swiss National Science Foundation
  • Deutsche Forschungsgemeinschaft
  • Klaus Tschira Stiftung

Zitieren

APA

Walz, E., Henzi, A., Ziegel, J., & Gneiting, T. (2024). Easy Uncertainty Quantification (EasyUQ): Generating Predictive Distributions from Single-Valued Model Output. SIAM Review, 66(1), 91-122. https://doi.org/10.1137/22m1541915

Vancouver

Walz E, Henzi A, Ziegel J, Gneiting T. Easy Uncertainty Quantification (EasyUQ): Generating Predictive Distributions from Single-Valued Model Output. SIAM Review. 2024;66(1):91-122. doi:10.1137/22m1541915

BibTeX
@article{Walz2024,
  author = {Walz, Eva-Maria and Henzi, Alexander and Ziegel, Johanna and Gneiting, Tilmann},
  title = {Easy Uncertainty Quantification (EasyUQ): Generating Predictive Distributions from Single-Valued Model Output},
  journal = {SIAM Review},
  year = {2024},
  volume = {66},
  number = {1},
  pages = {91--122},
  doi = {10.1137/22m1541915}
}
Kostenlos starten →