• Simone Pezzuto, Universitá Della Svizzera Italiana
• Francisco Sahli Costabal
• Rolf Krause
• Hermenegild Arevalo
• Luca Dedé

The human heart is amongst the most complicated biological systems to model. Cardiac models are genuinely multi-scale, multi-physics, and often hard to reduce to simple set of constitutive and conservation laws. Such complexity translates in a high computational burden in simulating realistic pathological scenarios as, for example, arrhythmic events, which in turn impedes robust parameter estimation and uncertainty quantification. Fortunately, the advance of artificial intelligence has widened the spectrum of techniques to cope with model complexity, providing new instruments in the toolbox of researchers. Machine learning not only can help to drastically reduce the computational cost of simulating established mathematical models; it can also unveil hidden physics or correlations in high-dimensional data and automatically translate them into effective low-dimensional models. The purpose of this mini symposium is, therefore, to bring together experts in cardiac modeling and machine learning for a fruitful discussion on how to blend diverse approaches for the common objective of simulating complex biological systems and enabling their practical employment. Oral contributions in the context of simulation-augmented machine learning, physics-informed learning, neural network model reduction, multi-fidelity techniques, uncertainty quantification applied to simulation of cardiac mechanics, electrophysiology and arrhythmias, and cardiac remodelling are welcome to be hosted in the mini symposium.