Efficient Medical VIE via Reinforcement Learning

Lijun Liu; Ruiyang Li; Zhaocheng Liu; Chenglin Zhu; Chong Li; Jiehan Cheng; Qiang Ju; Jian Xie

返回文献探索

Paper · arXiv 2506.13363

Efficient Medical VIE via Reinforcement Learning

Lijun Liu, Ruiyang Li, Zhaocheng Liu, Chenglin Zhu, Chong Li, Jiehan Cheng, Qiang Ju, Jian Xie

31 upvotesJune 16, 2025arXiv 预印本

AI 摘要

An RLVR framework using fine-tuned Qwen2.5-VL-7B achieves state-of-the-art performance in medical VIE with limited annotated samples, enhancing reasoning and balance between precision and recall.

Reinforcement Learning with Verifiable Rewards (RLVR)JSON generationmultimodal modelsdataset diversityprecision-recall reward mechanismhallucinationsfield coveragesampling strategiesfine-tuningQwen2.5-VL-7BF1 scorecase studies

Abstract

Visual Information Extraction (VIE) converts unstructured document images into structured formats like JSON, critical for medical applications such as report analysis and online consultations. Traditional methods rely on OCR and language models, while end-to-end multimodal models offer direct JSON generation. However, domain-specific schemas and high annotation costs limit their effectiveness in medical VIE. We base our approach on the Reinforcement Learning with Verifiable Rewards (RLVR) framework to address these challenges using only 100 annotated samples. Our approach ensures dataset diversity, a balanced precision-recall reward mechanism to reduce hallucinations and improve field coverage, and innovative sampling strategies to enhance reasoning capabilities. Fine-tuning Qwen2.5-VL-7B with our RLVR method, we achieve state-of-the-art performance on medical VIE tasks, significantly improving F1, precision, and recall. While our models excel on tasks similar to medical datasets, performance drops on dissimilar tasks, highlighting the need for domain-specific optimization. Case studies further demonstrate the value of reasoning during training and inference for VIE.