Offline Deep Reinforcement Learning and Off-Policy Evaluation for Personalized Basal Insulin Control in Type 1 Diabetes.
Taiyu ZhuKezhi LiPantelis GeorgiouPublished in: IEEE J. Biomed. Health Informatics (2023)
Keyphrases
- importance sampling
- policy evaluation
- monte carlo
- temporal difference
- reinforcement learning
- diabetes mellitus
- diabetic patients
- model free
- function approximation
- td learning
- least squares
- control system
- reinforcement learning algorithms
- optimal control
- blood glucose
- policy iteration
- machine learning
- fixed point
- state space
- e learning
- long run
- control strategy
- optical flow
- multi agent