基于声明式奖励优化长文本临床文本生成
Optimizing Long-Form Clinical Text Generation with Claim-Based Rewards
AI总结：
本研究提出集成GRPO与DocLens声明级评估器的强化学习框架，无需额外奖励模型或人工参考即可优化长临床文本的事实性与完整性，经实验和GPT-5评估验证有效，且可扩展适配真实场景自定义目标。
摘要：
利用大语言模型实现临床文档自动化需要与完整性、事实 grounded 性等优先级精准对齐。我们提出一种集成评估的强化学习框架用于长文本临床文本生成，该框架将Group Relative Policy Optimization（GRPO，组相对策略优化）与DocLens相结合，DocLens是一种声明级评估器，可提供确定性的、基于对话的奖励。我们的方法直接优化事实 grounded 性和完整性，无需训练单独的奖励模型，也不依赖人工撰写的参考文本。实验表明，该方法通过简单的奖励门控策略提升了临床笔记质量，同时降低了训练成本。独立的GPT-5定性评估进一步支持了这些提升，结果显示GRPO输出在事实性、完整性和简洁性上更受偏好，遗漏和幻觉更少。由于基准数据集相对干净且基础模型已经对齐良好，这些提升可能代表了保守的下限。该框架可扩展到真实场景，还可纳入自定义目标，如指南依从性或计费偏好。
英文摘要：
Automating clinical documentation with large language models requires precise alignment with priorities such as completeness and factual grounding. We present an evaluation-integrated reinforcement learning framework for long-form clinical text generation that couples Group Relative Policy Optimization (GRPO) with DocLens, a claim-level evaluator that provides deterministic, dialogue-grounded rewards. Our method directly optimizes factual grounding and completeness without training a separate reward model or relying on human-authored references. Empirically, the approach improves clinical note quality and reduces training cost via a simple reward-gating strategy. An independent GPT-5 qualitative evaluation further supports these gains, showing higher preference for GRPO outputs in factuality, completeness, and brevity, with fewer omissions and hallucinations. Because the benchmarks are relatively clean and the base model already well aligned, these improvements likely represent a conservative lower bound. The framework is scalable to real-world settings and can incorporate custom objectives such as guideline adherence or billing preferences.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
