Skip to main content
. 2020 May 6;15(5):e0232683. doi: 10.1371/journal.pone.0232683

Fig 4. Neural network architecture used for nonlinear diffusivity equation.

Fig 4

This network corresponds to the top part of Fig 3. There are two input nodes, x and t, and one output node, p. The number of hidden layers, Nhl, and the number of neurons for each hidden layer, Nn, denote the hyperparameters.