A View on Model Misspecification in Uncertainty Quantification

Yuko Kato*, David M.J. Tax, Marco Loog

*Corresponding author for this work

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

Abstract

Estimating uncertainty of machine learning models is essential to assess the quality of the predictions that these models provide. However, there are several factors that influence the quality of uncertainty estimates, one of which is the amount of model misspecification. Model misspecification always exists as models are mere simplifications or approximations to reality. The question arises whether the estimated uncertainty under model misspecification is reliable or not. In this paper, we argue that model misspecification should receive more attention, by providing thought experiments and contextualizing these with relevant literature.

Original languageEnglish
Title of host publicationArtificial Intelligence and Machine Learning - 34th Joint Benelux Conference, BNAIC/Benelearn 2022, Revised Selected Papers
EditorsToon Calders, Bart Goethals, Celine Vens, Jefrey Lijffijt
PublisherSpringer
Pages65-77
Number of pages13
ISBN (Print)9783031391439
DOIs
Publication statusPublished - 2023
EventArtificial Intelligence and Machine Learning - 34th Joint Benelux Conference, BNAIC/Benelearn 2022, Revised Selected Papers - Mechelen, Belgium
Duration: 7 Nov 20229 Nov 2022

Publication series

NameCommunications in Computer and Information Science
Volume1805 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

ConferenceArtificial Intelligence and Machine Learning - 34th Joint Benelux Conference, BNAIC/Benelearn 2022, Revised Selected Papers
Country/TerritoryBelgium
CityMechelen
Period7/11/229/11/22

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

  • Epistemic and Aleatoric uncertainty
  • Model misspecification
  • Uncertainty quantification

Fingerprint

Dive into the research topics of 'A View on Model Misspecification in Uncertainty Quantification'. Together they form a unique fingerprint.

Cite this