Guardado en:
Detalles Bibliográficos
Autores principales: Schaper, Ben, Di Folco, Maxime, Kainz, Bernhard, Schnabel, Julia A., Bercea, Cosmin I.
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2601.14827
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909996794511360
author Schaper, Ben
Di Folco, Maxime
Kainz, Bernhard
Schnabel, Julia A.
Bercea, Cosmin I.
author_facet Schaper, Ben
Di Folco, Maxime
Kainz, Bernhard
Schnabel, Julia A.
Bercea, Cosmin I.
contents Vision-Language Models show strong zero-shot performance for chest X-ray classification, but standard flat metrics fail to distinguish between clinically minor and severe errors. This work investigates how to quantify and mitigate abstraction errors by leveraging medical taxonomies. We benchmark several state-of-the-art VLMs using hierarchical metrics and introduce Catastrophic Abstraction Errors to capture cross-branch mistakes. Our results reveal substantial misalignment of VLMs with clinical taxonomies despite high flat performance. To address this, we propose risk-constrained thresholding and taxonomy-aware fine-tuning with radial embeddings, which reduce severe abstraction errors to below 2 per cent while maintaining competitive performance. These findings highlight the importance of hierarchical evaluation and representation-level alignment for safer and more clinically meaningful deployment of VLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2601_14827
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Measuring and Aligning Abstraction in Vision-Language Models with Medical Taxonomies
Schaper, Ben
Di Folco, Maxime
Kainz, Bernhard
Schnabel, Julia A.
Bercea, Cosmin I.
Artificial Intelligence
Vision-Language Models show strong zero-shot performance for chest X-ray classification, but standard flat metrics fail to distinguish between clinically minor and severe errors. This work investigates how to quantify and mitigate abstraction errors by leveraging medical taxonomies. We benchmark several state-of-the-art VLMs using hierarchical metrics and introduce Catastrophic Abstraction Errors to capture cross-branch mistakes. Our results reveal substantial misalignment of VLMs with clinical taxonomies despite high flat performance. To address this, we propose risk-constrained thresholding and taxonomy-aware fine-tuning with radial embeddings, which reduce severe abstraction errors to below 2 per cent while maintaining competitive performance. These findings highlight the importance of hierarchical evaluation and representation-level alignment for safer and more clinically meaningful deployment of VLMs.
title Measuring and Aligning Abstraction in Vision-Language Models with Medical Taxonomies
topic Artificial Intelligence
url https://arxiv.org/abs/2601.14827