Research
Jan Hagnberger develops neural surrogate models for time-dependent partial differential equations and aerodynamic simulation. His work includes conditional neural fields, efficient simulation in compressed latent spaces, and mesh-free models that predict flow fields directly from geometry. These methods aim to reduce simulation costs while accommodating different spatial discretizations and query locations.
Selected publications
- SMART: Scalable Mesh-free Aerodynamic Simulations from Raw Geometries using a Transformer-based Surrogate Model 2026
- CALM-PDE: Continuous and Adaptive Convolutions for Latent Space Modeling of Time-dependent PDEs 2025
- Vectorized Conditional Neural Fields: A Framework for Solving Time-dependent Parametric Partial Differential Equations 2024
Information reviewed 26 September 2026 · University profile