of the Austrian Academy of Sciences
“New Mathematical Tools developed in RICAM for Diabetes Technology”
We are going to present a new scheme of regularization algorithms, which is based on the concept of meta-learning and presupposes that the instances of algorithms, such as regularization spaces and parameters, are adaptively chosen on the base of the experience with similar problems. The construction of such fully adaptive regularization algorithms is motivated by the problem of predicting the blood glucose concentration of diabetic patients. We describe how proposed scheme can be used for this purpose.
The scheme is implemented in the device and tested in clinical trials performed within EU-project DIAdvisor. The presentation is based on the patent application WO 2012/143505 A2 filed jointly by Austrian Academy of Sciences and Novo Nordisk A/S (Denmark)