Second-Order Uncertainty Quantification: Variance-Based Measures
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914625927249920 |
|---|---|
| author | Sale, Yusuf Hofman, Paul Wimmer, Lisa Hüllermeier, Eyke Nagler, Thomas |
| author_facet | Sale, Yusuf Hofman, Paul Wimmer, Lisa Hüllermeier, Eyke Nagler, Thomas |
| contents | Uncertainty quantification is a critical aspect of machine learning models, providing important insights into the reliability of predictions and aiding the decision-making process in real-world applications. This paper proposes a novel way to use variance-based measures to quantify uncertainty on the basis of second-order distributions in classification problems. A distinctive feature of the measures is the ability to reason about uncertainties on a class-based level, which is useful in situations where nuanced decision-making is required. Recalling some properties from the literature, we highlight that the variance-based measures satisfy important (axiomatic) properties. In addition to this axiomatic approach, we present empirical results showing the measures to be effective and competitive to commonly used entropy-based measures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_00276 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Second-Order Uncertainty Quantification: Variance-Based Measures Sale, Yusuf Hofman, Paul Wimmer, Lisa Hüllermeier, Eyke Nagler, Thomas Machine Learning Uncertainty quantification is a critical aspect of machine learning models, providing important insights into the reliability of predictions and aiding the decision-making process in real-world applications. This paper proposes a novel way to use variance-based measures to quantify uncertainty on the basis of second-order distributions in classification problems. A distinctive feature of the measures is the ability to reason about uncertainties on a class-based level, which is useful in situations where nuanced decision-making is required. Recalling some properties from the literature, we highlight that the variance-based measures satisfy important (axiomatic) properties. In addition to this axiomatic approach, we present empirical results showing the measures to be effective and competitive to commonly used entropy-based measures. |
| title | Second-Order Uncertainty Quantification: Variance-Based Measures |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2401.00276 |