GFocal, a global-focal Transformer neural operator, achieves state-of-the-art relative L2 errors on five of six PDE benchmarks by fusing Nystrom attention and slice-based local tokens.
Nature machine intelligence, 3(3): 218–229
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
GFocal: A Global-Focal Neural Operator for Solving PDEs on Arbitrary Geometries
GFocal, a global-focal Transformer neural operator, achieves state-of-the-art relative L2 errors on five of six PDE benchmarks by fusing Nystrom attention and slice-based local tokens.