FNO with post-training least-squares readout refit outperforms vanilla FNO, DeepONet variants on random obstacle-to-solution maps from variational inequalities, especially for complex contact geometry.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.NA 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Fourier Neural Operators with Least-Squares Readout Refit for Learning Random Obstacle-to-Solution Maps
FNO with post-training least-squares readout refit outperforms vanilla FNO, DeepONet variants on random obstacle-to-solution maps from variational inequalities, especially for complex contact geometry.