Rethinking Aleatoric and Epistemic Uncertainty
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866911108356374528 |
|---|---|
| author | Smith, Freddie Bickford Kossen, Jannik Trollope, Eleanor van der Wilk, Mark Foster, Adam Rainforth, Tom |
| author_facet | Smith, Freddie Bickford Kossen, Jannik Trollope, Eleanor van der Wilk, Mark Foster, Adam Rainforth, Tom |
| contents | The ideas of aleatoric and epistemic uncertainty are widely used to reason about the probabilistic predictions of machine-learning models. We identify incoherence in existing discussions of these ideas and suggest this stems from the aleatoric-epistemic view being insufficiently expressive to capture all the distinct quantities that researchers are interested in. To address this we present a decision-theoretic perspective that relates rigorous notions of uncertainty, predictive performance and statistical dispersion in data. This serves to support clearer thinking as the field moves forward. Additionally we provide insights into popular information-theoretic quantities, showing they can be poor estimators of what they are often purported to measure, while also explaining how they can still be useful in guiding data acquisition. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_20892 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Rethinking Aleatoric and Epistemic Uncertainty Smith, Freddie Bickford Kossen, Jannik Trollope, Eleanor van der Wilk, Mark Foster, Adam Rainforth, Tom Machine Learning The ideas of aleatoric and epistemic uncertainty are widely used to reason about the probabilistic predictions of machine-learning models. We identify incoherence in existing discussions of these ideas and suggest this stems from the aleatoric-epistemic view being insufficiently expressive to capture all the distinct quantities that researchers are interested in. To address this we present a decision-theoretic perspective that relates rigorous notions of uncertainty, predictive performance and statistical dispersion in data. This serves to support clearer thinking as the field moves forward. Additionally we provide insights into popular information-theoretic quantities, showing they can be poor estimators of what they are often purported to measure, while also explaining how they can still be useful in guiding data acquisition. |
| title | Rethinking Aleatoric and Epistemic Uncertainty |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2412.20892 |