Lightweight Geometric Adaptation for Training Physics-Informed Neural Networks

Kang An, Chenhao Si, Shiqian Ma +1

Physics-Informed Neural Networks (PINNs) often suffer from slow convergence, training instability, and reduced accuracy on challenging partial differential equations due to the ani…