Researchers develop AI neural networks to identify blood vessel properties from pressure measurements, advancing non-invasive cardiovascular modeling

Uses Asymptotic-Preserving Neural Networks to embed physical principles of multiscale blood flow models into machine learning algorithms

Addresses the challenge of determining viscoelastic parameters that control how arterial walls deform under pulsatile pressure

Enables simultaneous inference of vessel parameters and reconstruction of time-dependent blood vessel state variables

Allows pressure waveforms to be estimated from readily accessible measurements, improving practical applicability of cardiovascular simulations