文献类型：期刊文献

英文题名：Enhancing Reinforcement Learning for Radiology Report Generation with Evidence-aware Rewards and Self-correcting Preference Learning

作者：Zhou, Qin[1,2]; Liang, Guoyan[3,4]; Yang, Qianyi[3,4]; Chen, Jingyuan[3,4]; Wu, Sai[3,4]; Yao, Chang[3,4]; Wang, Zhe[1,2]

机构：[1] Department of Computer Science and Engineering, ECUST, China; [2] Key Laboratory of Smart Manufacturing in Energy Chemical Process, Ministry of Education, China; [3] Zhejiang University, Hangzhou, China; [4] Hangzhou High-Tech Zone [Binjiang], Institute of Blockchain and Data Security, China

年份：2026

外文期刊名：arXiv

收录：EI(收录号：20260228037)

语种：英文

外文关键词：Cardiology - Electric grounding - Personnel training - Radiology

摘要：Recent reinforcement learning (RL) approaches have advanced radiology report generation (RRG), yet two core limitations persist: (1) report-level rewards offer limited evidence-grounded guidance for clinical faithfulness; and (2) current methods lack an explicit self-improving mechanism to align with clinical preference. We introduce clinically aligned Evidence-aware Self-Correcting Reinforcement Learning (ESC-RL), comprising two key components. First, a Group-wise Evidence-aware Alignment Reward (GEAR) delivers group-wise, evidence-aware feedback. GEAR reinforces consistent grounding for true positives, recovers missed findings for false negatives, and suppresses unsupported content for false positives. Second, a Self-correcting Preference Learning (SPL) strategy automatically constructs a reliable, disease-aware preference dataset from multiple noisy observations and leverages an LLM to synthesize refined reports without human supervision. ESC-RL promotes clinically faithful, disease-aligned reward and supports continual self-improvement during training. Extensive experiments on two public chest X-ray datasets demonstrate consistent gains and state-of-the-art performance. ? 2026, CC BY-NC-ND.