What does a Heart Beat for? A Heterogeneous Approach for Human Digital Twin Construction

Abstract

As healthcare applications of Artificial Intelligence are growing more and more, the necessity of building universal and transparent models in healthcare is of a paramount importance. Defining approaches able to combine the flexibility of data-driven and the explainability of model-based methods is still an open and promising challenge. This paper proposes an approach to combine mathematical modelling, based on Ordinary Differential Equations, Machine Learning and Fluid Stochastic Petri Nets. By adopting the proposed method and tools, a user could understand the effects of drugs and other treatments on a patient in an interpretable manner. The paper also presents a preliminary application of the approach on the human heartbeat.

Topics & Concepts