Full Professor, Swiss Federal Institute of Technology
3 papers at NeurIPS 2025
We propose an accurate and efficient neural operator architecture for learning PDE solutions on arbitrary domains. We demonstrate its effectiveness across a variety of challenging benchmarks, including large-scale 3D CFD problems.
We propose a very performant graph-based neural operator architecture for learning the solution operator of PDEs from data on arbitrary domain discretizations, which has been tested on a challenging suite of benchmark datasets.
