DiabAgent advances glucose control with adaptive AI decision tools
DiabAgent: Toward Agentic Physiological Control
Artificial IntelligenceComputation and Language
Summary
Managing blood sugar in type 1 diabetes requires carefully timed insulin doses based on many factors like meals and exercise. The paper presents DiabAgent, a new system that uses artificial intelligence to make smarter insulin dosing decisions by learning from continuous data about an individual’s condition. It includes a simulator to test how well different approaches work and shows that combining AI decision-making with rule-based control improves managing glucose levels. This work provides a way to build and test AI agents that could help control human physiology more effectively.
What this means in practice
- •For medical device developers: Test AI-driven insulin dosing strategies in a realistic simulator to improve automated glucose control devices.
- •For hospital diabetes care teams: Use adaptive AI frameworks to personalize insulin treatment plans based on patient-specific data and events.
Authors
Chengxi Li, Shaoxuan Li, Deyao Hong, Qingle Liu, Yi Huang, Puheng Bian, Minzhi Xu, Lanyue Zhang, Mingze Sun, Xin Li, Calvin Xiao, Shengyuan Wang, Huan-ang Gao, Jiazheng Xu, Kaisen Yang, Hongning Wang
Abstract
As pretraining scales, medical language models have advanced rapidly in clinical question answering. Their ability to achieve effective physiological control through sustained intervention still lacks systematic quantitative evaluation. Type 1 diabetes provides a concrete setting: impaired insulin secretion creates dependence on exogenous insulin, with dosing continually adjusted using information about meals, exercise, illness, and glucose measurements. We introduce DiabAgent, a research suite for agentic physiological control comprising the Agentic Physiological Control (APC) framework, the DiabetSim simulator, and the SweetBench benchmark. APC controls glucose through medical-device tools, using ongoing observations to infer individual states, accumulate personal experience, and choose treatment actions. We build DiabetSim to simulate different patients, events, and device conditions for scalable virtual testing and training. Using DiabetSim, we construct SweetBench by instantiating 32 scenario designs as 212 glucose-control tasks. With Claude Opus 5.5 and a fixed decision interval of one hour, APC achieves a COGI of 73.5, compared with 71.2 for the open-source closed-loop baseline OpenAPS. Ablations further show that individual adaptation and contextual information improve glucose control. Combining high-frequency rule-based control with lower-frequency agent decisions shows potential for balancing control performance and computational cost. Our work provides a framework, simulator, and evaluation foundation for agents to participate in human physiological control.