Artificial Intelligence #machine learning#physics-informed
Geometry-Aware Neural Operator Cuts Simulation Time for Plate Structures from Hours to Milliseconds
Researchers propose MR-GVNO, a geometry-aware variational neural operator for Mindlin-Reissner plate problems. The model uses boundary point clouds and cross-attention to predict responses on irregular domains, achieving millisecond-level inference without labeled training data.
Jun 16, 2026 1 source