The source code for the paper L. Lu, R. Pestourie, W. Yao, Z. Wang, F. Verdugo, & S. G. Johnson. Physics-informed neural networks with hard constraints for inverse design. arXiv preprint arXiv:2102.04626, 2021.

The code depends on the deep learning package DeepXDE v0.9.1. If you want to use the latest DeepXDE, you need to modify the code.

• hPINN
• FDFD: Jupyter Notebook
• FEM: Jupyter Notebook

• hPINN

If you use this code for academic research, you are encouraged to cite the following paper:

@article{lu2021physics,
  title   = {Physics-informed neural networks with hard constraints for inverse design},
  author  = {Lu, Lu and Pestourie, Raphael and Yao, Wenjie and Wang, Zhicheng and Verdugo, Francesc and Johnson, Steven G},
  journal = {arXiv preprint arXiv:2102.04626},
  year    = {2021}
}

To get help on how to use the code, simply open an issue in the GitHub "Issues" section.