Retraining only the final layer of a PINN on a few high-loss collocation points from the adjacent validation interval, plus a learnable activation function, reduces extrapolation error on Allen-Cahn, KdV, and Burgers benchmarks, but the claimed 40-50% average reductions are only achieved when…
Title resolution pending
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
-
Improving physics-informed neural network extrapolation via transfer learning and adaptive activation functions
Retraining only the final layer of a PINN on a few high-loss collocation points from the adjacent validation interval, plus a learnable activation function, reduces extrapolation error on Allen-Cahn, KdV, and Burgers benchmarks, but the claimed 40-50% average reductions are only achieved when…