Model-Based Reinforcement Learning for Type 1 Diabetes Blood Glucose Control

Taku Yamagata*, Amid Ayobi, Aisling Ann O'Kane, Dmitri Katz, Katarzyna Stawarz, Paul Marshall, Peter A Flach, Raul Santos-Rodriguez

*Corresponding author for this work

Research output: Contribution to conferenceConference Paper


In this paper we investigate the use of model-based reinforcement learning to assist people with Type 1 Diabetes with insulin dose decisions. The proposed architecture consists of multiple Echo State Networks to predict blood glucose levels combined with Model Predictive Controller for planning. Echo State Network is a version of recurrent neural networks which allows us to learn long term dependencies in the input of time series data in an online manner. Additionally, we address the quantification of uncertainty for a more robust control. Here, we used ensembles of Echo State Networks to capture model (epistemic) uncertainty.
We evaluated the approach with the FDA-approved UVa/Padova Type 1 Diabetes simulator and compared the results against baseline algorithms such as Basal-Bolus controller and Deep Q-learning. The results suggest that the model-based reinforcement learning algorithm can perform equally or better than the baseline algorithms for the majority of virtual Type 1 Diabetes person profiles tested.
Original languageEnglish
Number of pages14
Publication statusAccepted/In press - 26 Jun 2020
EventSingular Problems for Healthcare Workshop at ECAI 2020 - online
Duration: 29 Aug 20208 Sep 2020


WorkshopSingular Problems for Healthcare Workshop at ECAI 2020
Internet address


  • type 1 diabetes
  • decision making support
  • reinforcement learning
  • echo state networks

Fingerprint Dive into the research topics of 'Model-Based Reinforcement Learning for Type 1 Diabetes Blood Glucose Control'. Together they form a unique fingerprint.

Cite this