EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Monte Carlo conformal prediction for quantifying uncertainty in radio galaxy classification under ambiguous ground truth

Alex Walls, James Barry, Devina Mohan, Anna M M Scaife

RAS Techniques and Instruments · 06.01.2026 · 5

DOI: 10.1093/rasti/rzag033 ↗

Kurz gefasst

Fachartikel in der Zeitschrift RAS Techniques and Instruments, online erschienen am 27.05.2026. Verfasst von Alex Walls und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 94 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Alan Turing Institute.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dramatically increasing data volumes are forcing astronomers to adopt automated methods for the identification and classification of astronomical objects. Although deep-learning models are often well-suited to this task, obtaining a measure of uncertainty on their predictions is challenging. Here we consider the suitability of Monte Carlo conformal prediction (MCCP) set size and confidence as measures of model uncertainty for the astronomical classification of radio galaxies. We demonstrate this approach using model predictions from a pre-trained radio galaxy foundation model, fine-tuned on a smaller set of labelled radio galaxies. We calibrate the MCCP by obtaining annotator-derived soft label distributions, i.e. probability distributions over classes instead of single class assignments, for each of these labelled radio galaxies and compare the resulting set sizes and confidence scores to predictive entropy measures for each galaxy obtained using a supervised Bayesian deep-learning model trained using Hamiltonian Monte Carlo. The comparison reveals only a weak correlation between the measures. We suggest that this indicates that MCCP and predictive entropy capture fundamentally different aspects of the total uncertainty, and conclude that the applicability of MCCP in this context is severely limited by the practical considerations associated with obtaining soft label distributions for specialized classifications.

Förderung

  • Alan Turing Institute

Zitieren

APA

Walls, A., Barry, J., Mohan, D., & Scaife, A. M. M. (2026). Monte Carlo conformal prediction for quantifying uncertainty in radio galaxy classification under ambiguous ground truth. RAS Techniques and Instruments, 5. https://doi.org/10.1093/rasti/rzag033

Vancouver

Walls A, Barry J, Mohan D, Scaife AMM. Monte Carlo conformal prediction for quantifying uncertainty in radio galaxy classification under ambiguous ground truth. RAS Techniques and Instruments. 2026;5. doi:10.1093/rasti/rzag033

BibTeX
@article{Walls2026,
  author = {Walls, Alex and Barry, James and Mohan, Devina and Scaife, Anna M M},
  title = {Monte Carlo conformal prediction for quantifying uncertainty in radio galaxy classification under ambiguous ground truth},
  journal = {RAS Techniques and Instruments},
  year = {2026},
  volume = {5},
  doi = {10.1093/rasti/rzag033}
}
Kostenlos starten →