There is no denying the fact that diabetes is on a steady rise worldwide. It is one of those permanent conditions that require lifetime care. To help manage diabetes, an artificial pancreas system, which is capable of blood sugar levels to introduce an appropriate amount of insulin into the blood, has now become smarter than ever. All the credit goes to AI learning.
A new research team has recently developed a reinforcement learning (RL) based AI algorithm that can calculate the amount of insulin needed by a diabetic patient and injects it automatically.
About Existing Artificial Pancreas
Patients, who are suffering from type 1 diabetes, must inject insulin daily. It is vital to check the amount of carbohydrates in the food ingested each time, calculate the appropriate amount of insulin needed by the body, and then inject the correct dosage before each meal. To date, the artificial pancreas systems available could help with the process. However, someone would still have to go through the hassle to input the meal intake in advance each time.
When AI Joined the Party
The research team has found an interesting way to eliminate this inconvenience. They have been able to add a pharmacological concept to reinforcement learning, more popularly known as the algorithm of AlphaGo. This method of AI algorithm achieved mean glucose of 124.72 mg/dL with a percentage time in the normal range of 89.56 percent. Even without having to input the meal intake, the algorithm demonstrated performance comparable to that of the conventional artificial pancreas.
The Extensive Reach of AlphaGo
According to a member of the research team, the fully automated artificial pancreas system is just like autonomous driving for the medical industry. The newly developed AI algorithm is capable of enabling fully automated blood sugar control without the further hassle of inputting any information regarding meals or exercise. They also expect this algorithm to be helpful in other drug-based treatments.
