no code implementations • 28 Mar 2024 • Shahed Rezaei, Shirko Faroughi, Mahdi Asgharzadeh, Ali Harandi, Gottfried Laschet, Stefanie Reese, Markus Apel
Our method, inspired by operator learning and the finite element method, demonstrates the ability to train without relying on data from other numerical solvers.
no code implementations • 4 Jan 2024 • Shahed Rezaei, Ahmad Moeineddin, Michael Kaliske, Markus Apel
We benchmark our methodology against the standard finite element method, demonstrating accurate yet faster predictions using the trained neural network for temperature and flux profiles.