• Example Multilayer Neural Network (MNN)
  • input in this problem is vector $\vec x = \{x_0,x_1,x_2,x_3\}$
  • In order to pass to second layer of three nodes we have to multiply by a weight matrices $W$
  • $W_{\_ 1}$ is a 4 x 3 matrix
    • This is because each column of the layer is multiplied by a column of the matrix to form the corresponding node of the next
    • There are multiple weight matrices, 1 for each transfer between layers
  • final prediction is $\hat y$