Unmasking the chameleons: A benchmark for out-of-distribution detection in medical tabular data.

Journal: International journal of medical informatics
Published Date:

Abstract

BACKGROUND: Machine Learning (ML) models often struggle to generalize effectively to data that deviates from the training distribution. This raises significant concerns about the reliability of real-world healthcare systems encountering such inputs known as out-of-distribution (OOD) data. These concerns can be addressed by real-time detection of OOD inputs. While numerous OOD detection approaches have been suggested in other fields - especially in computer vision - it remains unclear whether similar methods effectively address challenges posed by medical tabular data.

Authors

  • Mohammad Azizmalayeri
    Department of Medical Informatics, Amsterdam Public Health Research Institute, Amsterdam UMC, University of Amsterdam, the Netherlands. Electronic address: m.azizmalayeri@amsterdamumc.nl.
  • Ameen Abu-Hanna
    Department of Medical Informatics, Amsterdam Public Health Research Institute, Amsterdam UMC, University of Amsterdam, Meibergdreef 9, 1105AZ, Amsterdam, The Netherlands.
  • Giovanni CinĂ 
    Pacmed B.V., Amsterdam, The Netherlands.