2

Let's consider this equation:

with : 0 < b < 1.

I am new to julia programming, I am considering to solve this ODE equation using NeuralPDE.jl. Any idea how to write that physics informed neural network ode in julia framework ?

logankilpatrick
  • 13,148
  • 7
  • 44
  • 125
Minou92
  • 137
  • 7
  • did you try coding up the model and see if it just works? I would think this example just works without doing anything fancy? – Chris Rackauckas Jan 03 '21 at 20:53
  • Thank you Chris for your reply. In fact, I have been trying to implement Maziar Raissi code at first, then I came across NeuralPDE.jl and DiffEqFlux.jl which I found very interesting. The problem is that we have a time series data (non-stationary), then we use analytical solution to fit the curve for foracasting,. The idea is to use the ODE equation to regularize the fitting. The parameter K and b are equation parameters and evaluated at the same time as the neural network’s weights. I have just started learning Julia, and I will try to build the model. – Minou92 Jan 03 '21 at 21:52

0 Answers0