Reinforcement Learning Improves LLM Accuracy and Reasoning in Disease Classification from Radiology Reports
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913050866483200 |
|---|---|
| author | Wei, Yishu Lin, Yi Flanders, Adam Shih, George Peng, Yifan |
| author_facet | Wei, Yishu Lin, Yi Flanders, Adam Shih, George Peng, Yifan |
| contents | Accurate disease classification from radiology reports is essential for many applications. While supervised fine-tuning (SFT) of lightweight LLMs improves accuracy, it can degrade reasoning. We propose a two-stage approach: SFT on disease labels followed by Group Relative Policy Optimization (GRPO) to refine predictions by optimizing accuracy and format without reasoning supervision. Across three radiologist-annotated datasets, SFT outperformed baselines and GRPO further improved classification and enhanced reasoning recall and comprehensiveness. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_19060 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Reinforcement Learning Improves LLM Accuracy and Reasoning in Disease Classification from Radiology Reports Wei, Yishu Lin, Yi Flanders, Adam Shih, George Peng, Yifan Artificial Intelligence Accurate disease classification from radiology reports is essential for many applications. While supervised fine-tuning (SFT) of lightweight LLMs improves accuracy, it can degrade reasoning. We propose a two-stage approach: SFT on disease labels followed by Group Relative Policy Optimization (GRPO) to refine predictions by optimizing accuracy and format without reasoning supervision. Across three radiologist-annotated datasets, SFT outperformed baselines and GRPO further improved classification and enhanced reasoning recall and comprehensiveness. |
| title | Reinforcement Learning Improves LLM Accuracy and Reasoning in Disease Classification from Radiology Reports |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2604.19060 |