Reinforcement Learning Significantly Outperforms Commercial Blood Glucose Controllers

Scientists at the University of Bristol have demonstrated that reinforcement learning, a type of machine learning, can be used to help type-1 diabetes sufferers keep their blood glucose levels in a safe range. The study, published in the Journal of Biomedical Informatics, showed that offline reinforcement learning significantly outperforms commercial blood glucose controllers in terms of safety and effectiveness. The largest improvement was in children, who experienced an additional one-and-a-half hours in the target glucose range per day. This research explores the potential of reinforcement learning to develop safer and more effective insulin dosing strategies.