A two-stage PINN uses Fourier series for temperature and energy-loss MLP for warpage to match FEM results at 0.2 μm MAE with 1000x speedup on CTE studies.
Physics- informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,
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
-
WarpagePINN: Thermal Warpage Prediction in Advanced Packaging via a Two-Stage Physics-Informed Neural Networks
A two-stage PINN uses Fourier series for temperature and energy-loss MLP for warpage to match FEM results at 0.2 μm MAE with 1000x speedup on CTE studies.