lululxvi / deepxde

A library for scientific machine learning and physics-informed learning
https://deepxde.readthedocs.io
GNU Lesser General Public License v2.1
2.73k stars 756 forks source link

Multi-fidelity fusion of physical information #590

Closed LanPeng-94 closed 2 years ago

LanPeng-94 commented 2 years ago

Dear Dr.Lu: After I examined Deepxde's source code and the case of mf presented, I found no mention of how to incorporate physical information. As shown in the figure below, the ideas mentioned by Dr. Meng, the loss term MSEfe of physical information can be added to the total loss function. However, in Deepxde, the MfFunc function class seems to mention only the mathematical functions that incorporate multi-fidelity and low-fidelity, and not how to incorporate physical information. image

1649062388(1)

haison19952013 commented 2 years ago

Physics-informed MFNN has not been implemented in deepXDE.. @979736316 .. image

LanPeng-94 commented 2 years ago

Physics-informed MFNN has not been implemented in deepXDE

Thank you for your reply, I look forward to the development of this feature.