This article was accepted into the Deep Generative Models for Health workshop at NeurIPS 2023.
Cardiovascular diseases (CVD) are a major global health problem, making longitudinal monitoring of cardiovascular biomarkers vital for early diagnosis and intervention. A central challenge is the inference of cardiac pulse parameters from pulse waves, especially when acquired from wearable sensors at peripheral locations on the body. Traditional machine learning (ML) approaches face obstacles in this context due to the scarcity of labeled data, mainly coming from clinical settings. At the same time, physical models such as whole-body 1D hemodynamics simulators, although informative, struggle with the inverse problem and complications posed by parameter interactions. Recent work has turned to simulation-based inference (SBI) to inform parameter inference leveraging model simulations. Still, transferring predictors from simulations to real-world data remains challenging due to model misspecifications. Addressing these issues, this article presents a novel hybrid learning approach. By fusing a pulse wave propagation simulator with a data-driven correction model, our methodology aims to combine the rigor of physical models with the flexibility of ML, offering a promising avenue for effective cardiovascular biomarker monitoring.
