Saved in:
Bibliographic Details
Main Authors: Li, Jun, Liu, Mingxuan, Pan, Jiazhen, Liu, Che, Bai, Wenjia, Bercea, Cosmin I., Schnabel, Julia A.
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.24972
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913067015602176
author Li, Jun
Liu, Mingxuan
Pan, Jiazhen
Liu, Che
Bai, Wenjia
Bercea, Cosmin I.
Schnabel, Julia A.
author_facet Li, Jun
Liu, Mingxuan
Pan, Jiazhen
Liu, Che
Bai, Wenjia
Bercea, Cosmin I.
Schnabel, Julia A.
contents Clinical abnormality grounding for rare diseases is often hindered by data scarcity, making supervised fine-tuning impractical and single-pass inference highly unstable. We propose Dynamic Decision Learning (DDL), a framework that enables frozen large vision-language models (LVLMs) to refine their decisions across both language and visual spaces by optimizing instructions and consolidating predictions under visual perturbations. This process improves localization quality and produces a consensus-based reliability score that quantifies model confidence. Results on brain imaging benchmarks, including a rare-disease dataset with 281 pathology types across models ranging from 3B to 72B parameters, show that DDL improves mAP@75 by up to 105% on rare-disease cases and outperforms adaptation baselines and supervised fine-tuning. Furthermore, DDL demonstrates stronger calibration between reliability scores and localization accuracy under severe distribution shifts and increasing task difficulty. Code is available at: https://lijunrio.github.io/DDL/
format Preprint
id arxiv_https___arxiv_org_abs_2604_24972
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Dynamic Decision Learning: Test-Time Evolution for Abnormality Grounding in Rare Diseases
Li, Jun
Liu, Mingxuan
Pan, Jiazhen
Liu, Che
Bai, Wenjia
Bercea, Cosmin I.
Schnabel, Julia A.
Computation and Language
Clinical abnormality grounding for rare diseases is often hindered by data scarcity, making supervised fine-tuning impractical and single-pass inference highly unstable. We propose Dynamic Decision Learning (DDL), a framework that enables frozen large vision-language models (LVLMs) to refine their decisions across both language and visual spaces by optimizing instructions and consolidating predictions under visual perturbations. This process improves localization quality and produces a consensus-based reliability score that quantifies model confidence. Results on brain imaging benchmarks, including a rare-disease dataset with 281 pathology types across models ranging from 3B to 72B parameters, show that DDL improves mAP@75 by up to 105% on rare-disease cases and outperforms adaptation baselines and supervised fine-tuning. Furthermore, DDL demonstrates stronger calibration between reliability scores and localization accuracy under severe distribution shifts and increasing task difficulty. Code is available at: https://lijunrio.github.io/DDL/
title Dynamic Decision Learning: Test-Time Evolution for Abnormality Grounding in Rare Diseases
topic Computation and Language
url https://arxiv.org/abs/2604.24972