用大模型让糖尿病胰岛素泵既精准又可解释。
Interpretable Language Model for Closed-Loop Type 1 Diabetes Control
- 用强化学习训练专家系统，再将知识蒸馏到大模型中。
- 血糖控制达标率73.5%，且通过了严格安全验证。
- 适合关注医疗AI可解释性与安全性的研究者。
1型糖尿病是一种慢性、危及生命的自身免疫疾病，特征为胰腺β细胞完全破坏。尽管基于强化学习的人工胰腺系统在自动化胰岛素输注方面展现出潜力，但其“黑箱”特性使患者和医生难以充分信任。本文提出LLM-T1D，结合强化学习的精确性与大语言模型的人类化推理能力，构建更透明可靠的胰岛素泵控制器。通过训练专家级强化学习系统，并将其知识蒸馏至微调后的LLaMA 3.1 8B和Qwen3 8B模型，开发出性能超越原强化学习系统的控制器，同时能以通俗易懂的语言解释决策过程。在经FDA批准的UVA/Padova 1型糖尿病模拟器上测试，该模型实现73.5%的时间血糖在目标范围内（Time in Range），并满足严格的防幻觉形式化安全验证。
原文摘要 · Abstract (English)
Type 1 Diabetes (T1D) is a chronic, life-threatening autoimmune condition characterized by the complete destruction of insulin-producing pancreatic beta cells. While Artificial Pancreas Systems (APS) powered by Reinforcement Learning (RL) have shown promise in automating insulin delivery, their ``black-box'' nature makes it hard for patients and doctors to trust them fully. This paper presents LLM-T1D, a promising approach that combines the precision of RL with the clear, human-like reasoning of Large Language Models (LLMs) to create a more transparent and reliable insulin pump controller. By training an expert RL system and distilling its knowledge into fine-tuned LLaMA 3.1 8B and Qwen3 8B models, we developed a controller that not only surpasses the RL system's performance but also explains its decisions in plain, understandable language. Tested on the FDA-approved UVA/Padova T1D simulator, the LLM controllers deliver excellent blood sugar control (73.5% Time in Range) while maintaining strict formal safety verification against hallucinations.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。