Tag
The author trained an encoder-only transformer to predict future blood glucose levels from past glucose, carbs, insulin, and future meal/insulin inputs, releasing the MIT-licensed source code with pretrained weights.
Proposes EVIDENT, a framework that integrates Bayesian training and evidence-based ranking for neural architecture selection, demonstrated on subject-specific blood glucose forecasting in type 1 diabetes, systematically selecting low-capacity models that generalize reliably.