Hrvoje Abraham

Hrvoje Abraham graduated in theoretical physics from the University of Zagreb, after which he continued to expand his core research and engineering skills at the Ruđer Bošković Institute. He has been involved with software throughout his entire educational and professional life, having had the opportunity to work in the financial sector, lead a technology startup company, and currently serves as the Department Manager for Advanced Simulation Technologies at the Austrian technology giant AVL.

Simulation AI – Optimization is all you need

A novel branch of computational simulation technology emerged a few years ago and has been rapidly developed and utilized for increasingly complex problems. Physics Informed Neural Networks (PINNs) leverage the growing popularity and efficiency of GPU-parallelized machine learning models to provide a universal approach for solving numerical physical problems. This presentation is divided into three parts: an abstract and specific approach to reducing a problem to optimization, an explanation of the basic principles and examples behind PINNs, and taking an educated guess on the future development of this technology and its significance.