The authors developed a theoretical framework for learning operators from geometric manifolds to function spaces of physical fields.
Transformer maps CAD designs directly to physics fields, skipping meshing
A neural operator tokenizes NURBS patch parameters to predict continuous aerodynamic fields and enables gradient-based inverse design.
Big Tech
Daniel Leibovici · Nikola Borislavov Kovachki · Dawon Ahn · Ruben Ohana · Ira J. S. Shokar · Abouzar Ghasemi · +5 more
NVIDIA
Research Digest··2 min read
Leibovici et al.
Why this paper
From NVIDIA
In one line
CANTO predicts physical fields directly from continuous CAD geometry without meshing, achieving state-of-the-art accuracy on aerodynamics benchmarks.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§