ARLPE: A meta reinforcement learning framework for glucose regulation in type 1 diabetics

Xuehui Yu, Yi Guan, Lian Yan, Shulang Li, Xuelian Fu, Jingchi Jiang

Abstract: Highlights•Automatic glucose regulation of unknown patients with uncertainties and noise.•A glucose regulation framework that enables fast adaptation and efficient training.•The combination of active learning and meta-reinforcement learning.•Active learning module increases the utilization of experience.•Patients’ probabilistic embeddings help to generate personalized strategies.