HAD-Net: Hybrid Attention-based Diffusion Network for Glucose Level\n Forecast

Résumé fourni par la source

Data-driven models for glucose level forecast often do not provide meaningful\ninsights despite accurate predictions. Yet, context understanding in medicine\nis crucial, in particular for diabetes management. In this paper, we introduce\nHAD-Net: a hybrid model that distills knowledge into a deep neural network from\nphysiological models. It models glucose, insulin and carbohydrates diffusion\nthrough a biologically inspired deep learning architecture tailored with a\nrecurrent attention network constrained by ODE expert models. We apply HAD-Net\nfor glucose level forecast of patients with type-2 diabetes. It achieves\ncompetitive performances while providing plausible measurements of insulin and\ncarbohydrates diffusion over time.\n

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.