Credal ensembling in multi-class classification
Machine Learning · 2025-01 · 114 (1)
DOI: 10.1007/s10994-024-06703-y ↗Kurz gefasst
Fachartikel in der Zeitschrift Machine Learning, online erschienen am 16.01.2025. Verfasst von Vu-Linh Nguyen und 2 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 61 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Junior Professor Chair in Trustworthy AI, Université de Technologie de Compiègne.
Abstract (Originalsprache)
In this paper, we present a formal framework to (1) aggregate probabilistic ensemble members into either a representative classifier or a credal classifier, and (2) perform various decision tasks based on this uncertainty quantification. We first elaborate on the aggregation problem under a class of distances between distributions. We then propose generic methods to robustify uncertainty quantification and decisions, based on the obtained ensemble and representative probability. To facilitate the scalability of the proposed framework, for all the problems and applications covered, we elaborate on their computational complexities from the theoretical aspects and leverage theoretical results to derive efficient algorithmic solutions. Finally, relevant sets of experiments are conducted to assess the usefulness of the proposed framework in uncertainty sampling, classification with a reject option, and set-valued prediction-making.
Förderung
- Junior Professor Chair in Trustworthy AI
- Université de Technologie de Compiègne
Zitieren
APA
Nguyen, V., Zhang, H., & Destercke, S. (2025). Credal ensembling in multi-class classification. Machine Learning, 114(1). https://doi.org/10.1007/s10994-024-06703-y
Vancouver
Nguyen V, Zhang H, Destercke S. Credal ensembling in multi-class classification. Machine Learning. 2025;114(1). doi:10.1007/s10994-024-06703-y
BibTeX
@article{Nguyen2025,
author = {Nguyen, Vu-Linh and Zhang, Haifei and Destercke, Sébastien},
title = {Credal ensembling in multi-class classification},
journal = {Machine Learning},
year = {2025},
volume = {114},
number = {1},
doi = {10.1007/s10994-024-06703-y}
}
EUVIMED