Integrating Neural Differential Forecasting with Safe Reinforcement Learning for Blood Glucose Regulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Yushen, Zhang, Yanfu, Zhou, Xugui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911391273713664
author Liu, Yushen
Zhang, Yanfu
Zhou, Xugui
author_facet Liu, Yushen
Zhang, Yanfu
Zhou, Xugui
contents Automated insulin delivery for Type 1 Diabetes must balance glucose control and safety under uncertain meals and physiological variability. While reinforcement learning (RL) enables adaptive personalization, existing approaches struggle to simultaneously guarantee safety, leaving a gap in achieving both personalized and risk-aware glucose control, such as overdosing before meals or stacking corrections. To bridge this gap, we propose TSODE, a safety-aware controller that integrates Thompson Sampling RL with a Neural Ordinary Differential Equation (NeuralODE) forecaster to address this challenge. Specifically, the NeuralODE predicts short-term glucose trajectories conditioned on proposed insulin doses, while a conformal calibration layer quantifies predictive uncertainty to reject or scale risky actions. In the FDA-approved UVa/Padova simulator (adult cohort), TSODE achieved 87.9% time-in-range with less than 10% time below 70 mg/dL, outperforming relevant baselines. These results demonstrate that integrating adaptive RL with calibrated NeuralODE forecasting enables interpretable, safe, and robust glucose regulation.
format Preprint
id arxiv_https___arxiv_org_abs_2511_12417
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Integrating Neural Differential Forecasting with Safe Reinforcement Learning for Blood Glucose Regulation
Liu, Yushen
Zhang, Yanfu
Zhou, Xugui
Machine Learning
Automated insulin delivery for Type 1 Diabetes must balance glucose control and safety under uncertain meals and physiological variability. While reinforcement learning (RL) enables adaptive personalization, existing approaches struggle to simultaneously guarantee safety, leaving a gap in achieving both personalized and risk-aware glucose control, such as overdosing before meals or stacking corrections. To bridge this gap, we propose TSODE, a safety-aware controller that integrates Thompson Sampling RL with a Neural Ordinary Differential Equation (NeuralODE) forecaster to address this challenge. Specifically, the NeuralODE predicts short-term glucose trajectories conditioned on proposed insulin doses, while a conformal calibration layer quantifies predictive uncertainty to reject or scale risky actions. In the FDA-approved UVa/Padova simulator (adult cohort), TSODE achieved 87.9% time-in-range with less than 10% time below 70 mg/dL, outperforming relevant baselines. These results demonstrate that integrating adaptive RL with calibrated NeuralODE forecasting enables interpretable, safe, and robust glucose regulation.
title Integrating Neural Differential Forecasting with Safe Reinforcement Learning for Blood Glucose Regulation
topic Machine Learning
url https://arxiv.org/abs/2511.12417