Time-Dependent PDE and RNN

[KirillZubov]

https://arxiv.org/abs/2009.03892

[killah-t-cell]

@KirillZubov I skimmed through the paper and it is very interesting. Do you know what the authors mean by data? Does it just mean feeding in the boundary conditions and initial conditions with correct values? Or does it mean actually having experimental data (I think it is called solving a reverse problem – where you have the data and need to discover the PDE value)?

[KirillZubov]

Yeah, it is a sort of data-driven method in the article. The main point here is that LSTM should be the most suitable NN for predicting time-dependent PDE in generally.