Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Zhuoxiao, Yu, Hongyang, Xu, Ying, Luo, Yadan, Duong, Long, Li, Yuan-Fang
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2509.18600
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911513683427328
author Chen, Zhuoxiao
Yu, Hongyang
Xu, Ying
Luo, Yadan
Duong, Long
Li, Yuan-Fang
author_facet Chen, Zhuoxiao
Yu, Hongyang
Xu, Ying
Luo, Yadan
Duong, Long
Li, Yuan-Fang
contents Radiology report generation (RRG) aims to automatically produce clinically faithful reports from chest X-ray images. Prevailing work typically follows a scale-driven paradigm, by multi-stage training over large paired corpora and oversized backbones, making pipelines highly data- and compute-intensive. In this paper, we propose Oracle-educated GRPO (OraPO) with a FactScore-based reward (FactS) to tackle the RRG task under constrained budgets. OraPO enables single-stage, RL-only training by converting failed GRPO explorations on rare or difficult studies into direct preference supervision via a lightweight oracle step. FactS grounds learning in diagnostic evidence by extracting atomic clinical facts and checking entailment against ground-truth labels, yielding dense, interpretable sentence-level rewards. Together, OraPO and FactS create a compact and powerful framework that significantly improves learning efficiency on clinically challenging cases, setting the new SOTA performance on the CheXpert Plus dataset (0.341 in F1) with 2--3 orders of magnitude less training data using a small base VLM on modest hardware.
format Preprint
id arxiv_https___arxiv_org_abs_2509_18600
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OraPO: Oracle-educated Reinforcement Learning for Data-efficient and Factual Radiology Report Generation
Chen, Zhuoxiao
Yu, Hongyang
Xu, Ying
Luo, Yadan
Duong, Long
Li, Yuan-Fang
Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
Radiology report generation (RRG) aims to automatically produce clinically faithful reports from chest X-ray images. Prevailing work typically follows a scale-driven paradigm, by multi-stage training over large paired corpora and oversized backbones, making pipelines highly data- and compute-intensive. In this paper, we propose Oracle-educated GRPO (OraPO) with a FactScore-based reward (FactS) to tackle the RRG task under constrained budgets. OraPO enables single-stage, RL-only training by converting failed GRPO explorations on rare or difficult studies into direct preference supervision via a lightweight oracle step. FactS grounds learning in diagnostic evidence by extracting atomic clinical facts and checking entailment against ground-truth labels, yielding dense, interpretable sentence-level rewards. Together, OraPO and FactS create a compact and powerful framework that significantly improves learning efficiency on clinically challenging cases, setting the new SOTA performance on the CheXpert Plus dataset (0.341 in F1) with 2--3 orders of magnitude less training data using a small base VLM on modest hardware.
title OraPO: Oracle-educated Reinforcement Learning for Data-efficient and Factual Radiology Report Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2509.18600