AI News HubLIVE
原文2 min read

Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs

Medical LVLMs are prone to factual inconsistencies and poor visual grounding. Existing alignment methods have three key limitations in the medical domain: sequence-level rewards treat clinically critical tokens equally, reliance on static SFT references causes off-policy shift, and alignment lacks visual grounding constraints. The proposed method uses a bidirectional token-wise KL regularizer and a visual-contrastive grounding objective, forming a fine-grained on-policy alignment framework that constructs preference pairs by minimally editing model outputs. Experiments validate its effectiveness.

SourcearXiv Computer VisionAuthor: Shayan Mohammadizadehsamakosh, Pritam Sarkar, Leonid Sigal, Ali Etemad, Elham Dolatabadi

[2606.12590] Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs

[Submitted on 10 Jun 2026]

Title:Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs

View a PDF of the paper titled Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs, by Shayan Mohammadizadehsamakosh and 4 other authors

View PDF HTML (experimental)

Abstract:Large Vision-Language Models (LVLMs) have achieved strong performance across medical imaging tasks, yet they remain prone to factual inconsistencies, poor visual grounding, and misalignment with clinically meaningful feedback. Existing post-training alignment approaches, including Direct Preference Optimization (DPO) and its variants, face three critical limitations in the medical domain: (1) sequence-level reward signals treat clinically critical tokens identically to generic filler text; (2) reliance on static supervised fine-tuning references as preferred responses introduces an off-policy distribution shift, steering optimization toward stylistic artifacts over clinical correctness; and (3) alignment objectives lack explicit visual grounding constraints, leaving models insensitive to subtle yet diagnostically decisive pathological features. Our method leverages a bidirectional token-wise KL regularizer alongside a visual-contrastive grounding objective that pairs clean and lesion-corrupted images to penalize responses generated without adequate visual evidence. Together, these components form a fine-grained, on-policy alignment framework that constructs preference pairs by minimally editing model-generated outputs, correcting only clinically erroneous spans while preserving the original linguistic style. Extensive experiments across medical imaging tasks and clinical text generation benchmarks validate the effectiveness of our approach.

Subjects:

Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)

Cite as: arXiv:2606.12590 [cs.CV]

(or arXiv:2606.12590v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2606.12590

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Shayan Mohammadizadehsamakosh [view email] [v1] Wed, 10 Jun 2026 18:35:36 UTC (5,015 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Analyzing and Improving Fine-grained Preference Optimization in Medical LVLMs, by Shayan Mohammadizadehsamakosh and 4 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CV

new | recent | 2026-06

Change to browse by:

cs cs.AI

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)