Neural Networks and Transfer Learning techniques to forecast blood glucose prediction.
External reference persons Alessandro Aliberti (firstname.lastname@example.org)
Thesis type EXPERIMENTAL
Description Diabetes is an autoimmune disease characterized by glucose levels dysfunctions. It involves continuous monitoring combined with insulin treatment. Nowadays, continuous glucose monitoring systems (CGMs) have led to a greater availability of data. These can be effectively used by machine learning techniques to infer future values of the glycaemic concentration, allowing the early prevention of dangerous states and a better optimisation of the diabetic treatment.
This thesis aims at developing an innovative specialized prediction model, based on neural networks, that, originally trained with multi-patient data, it adapts to the needs of the Type I diabetes single-patient. In a nutshell, a universal algorithm to be equipped on CGMs i) ready to use and ii) self-adaptive to any patient profile.
By addressing the problem of automated glucose level prediction CGMs data, the student will combine our patient specialized data-driven system with our multi-patient data-driven methodology by integrating run-time information. More specifically, the student will perform a real-time fine-tuning of the model, leveraging the glucose level measurements of the patient that is currently using the system.
Required skills Python
Deadline 18/10/2020 PROPONI LA TUA CANDIDATURA