FALM-PINN alternates basis learning with Levenberg-Marquardt coefficient fitting to solve high-frequency and nonlinear PDEs with orders-of-magnitude better accuracy than baselines.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Alternating Levenberg-Marquardt Training of Physics-Informed Neural Networks with Fourier-Enhanced Features
FALM-PINN alternates basis learning with Levenberg-Marquardt coefficient fitting to solve high-frequency and nonlinear PDEs with orders-of-magnitude better accuracy than baselines.