Diabetes is on the rise around the world. It is a permanent condition that requires care throughout life. To help manage it, the artificial pancreas system, which automatically measures blood sugar levels to inject the right amount of insulin into the blood, has now become smarter thanks to learning AI.
The research team, led by Professor Sung-Min Park and others. candidate Seunghyun Lee and MS candidate Jiwon Kim from POSTECH’s Department of Convergence of IT Engineering and Electrical Engineering, recently developed an AI algorithm based on learning reinforcement (RL) that calculates the amount of insulin needed for a diabetic and injects it automatically.
These findings were published as articles in the latest issue of the journal IEEE Journal of Biomedical and Health Informatics, international journal of medical informatics.
Patients with type 1 diabetes must inject insulin daily. Each time, you should check the amount of carbohydrates in the food you eat, calculate the appropriate amount of insulin, and then inject the correct dose before each meal. Although the artificial pancreatic system on the market helps with this process, there is still trouble about having to eat a meal in advance each time.
To eliminate this inconvenience, the research team added a pharmacological concept to amplification learning, widely known as the AlphaGo algorithm. This method of the AI algorithm achieved a mean glucose value of 124.72 mg / dL and a percentage time in the normal range of 89.56%. Even without meal intake, the policy showed an effect comparable to the effect of a conventional artificial pancreas.
A fully automated artificial pancreas is like an autonomous drive for the medical industry. The newly developed AI algorithm allows for fully automated blood sugar control without the hassle of entering meal or exercise information. We expect this algorithm to be extended to other drug-based treatments. “
Sung-Min Park, Professor, POSTECH
Pohang University of Science and Technology (POSTECH)
Lee, S., and others. (2021) Towards a fully automated artificial pancreatic system using Bioinspired learning design for reinforcement: in Silico Validation. IEEE Journal of Biomedical and Health Informatics. doi.org/10.1109/JBHI.2020.3002022.