post-training-optimization

This paper proposes a Physics-Informed Error Field Learning (PIEFL) framework for Physics-Informed Neural Networks, introducing an auxiliary error network to improve solution accuracy in solving partial differential equations under computational constraints.