lu-group/deeponet-extrapolation

DeepONet extrapolation

Reliable extrapolation of deep neural operators informed by physics or sparse observations

The data and code for the paper M. Zhu, H. Zhang, A. Jiao, G. E. Karniadakis, & L. Lu. Reliable extrapolation of deep neural operators informed by physics or sparse observations. Computer Methods in Applied Mechanics and Engineering, 412, 116064, 2023.

Datasets

• Diffusion-reaction equation
• Burgers’ equation
• Advection equation
• Poisson equation
• Cavity flow

To get complete datasets, run data_gen.py and data_gen_low_fidelity.py in the data folder for each equation.

Code

• Diffusion-reaction equation
• Burgers’ equation
• Advection equation
• Poisson equation
• Cavity flow

Cite this work

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

@article{zhu2023reliable,
  title   = {Reliable extrapolation of deep neural operators informed by physics or sparse observations},
  author  = {Zhu, Min and Zhang, Handi and Jiao, Anran and Karniadakis, George Em and Lu, Lu},
  journal = {Computer Methods in Applied Mechanics and Engineering},
  volume  = {412},
  pages   = {116064},
  year    = {2023},
  doi     = {https://doi.org/10.1016/j.cma.2023.116064}
}

Questions

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