arXiv cs.CL· Xu Zhang, Wenxin Ma, Chenxu Wu, Rongsheng Wang, Zhiyang He, Xiaodong Tao, Kun Zhang, S. Kevin Zhou·· 2 天前AI 评分22
从文档到片段：面向 LLM 可解释医疗编码的片段中心学习
From Documents to Spans: Span-Centric Learning for LLM-based Explainable Medical Coding
AI 导读
研究者提出 Span-Centric Learning（SCL），将编码知识获取与在完整文档中的应用分离，先让模型从短片段学习编码知识，再用少量标注数据学会在完整文档中运用。实验显示，该方法仅用 248 份完整文档训练，就让一个 8B LLM 超越现有最优方法，并为每个预测编码提供与人工标注高度一致的显式支持证据。分析还表明，模型能通过片段学习文档中未出现的编码知识，并将其应用于完整文档。
正文 · 原文
Abstract:Medical coding, which assigns standardized ICD codes to clinical documents, is a fundamental yet challenging task in modern healthcare. It requires extensive knowledge of a large code space and the ability to identify and interpret sparse supporting evidence in long, noisy clinical documents. Most existing methods rely on document-level supervision, which provides limited guidance on supporting evidence and insufficient training signals for rare codes. To address these limitations, we propose Span-Centric Learning (SCL), which separates coding knowledge acquisition from its application in full documents. SCL first learns coding knowledge from short spans, then learns to use this knowledge in full documents from a small set of annotated data. Experiments show that, our approach enables an 8B LLM to surpass state-of-the-art methods using only 248 complete documents for training. It also provides explicit supporting evidence for each predicted code, with strong alignment to human annotations. Further analysis shows that the model can learn knowledge of codes absent from documents through spans and apply that knowledge to full documents.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2603.15270 [cs.CL] |
| (or arXiv:2603.15270v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2603.15270 arXiv-issued DOI via DataCite |
Submission history
From: Xu Zhang [view email]
[v1]
Mon, 16 Mar 2026 13:37:01 UTC (9,377 KB)
[v2]
Thu, 7 May 2026 13:12:33 UTC (446 KB)
[v3]
Wed, 14 Oct 2026 08:11:59 UTC (767 KB)
来源：arXiv cs.CL · arxiv.org