C1_W4.pdf

What is a deep neural network

Notation

4 Layer NN with 3 Hidden Layer

$L = 4$

$n^{[l]} =$ units in layer l

$n^{[0]} = n_{x}$ the number of inputs

$a^{[l]}$ = activations in layer l

Forward Propagation in a Deep Network

We just do the same but iterating L times

$a^{[0]} = [x_1,x_2,...,x_n]$ input layer

$z^{[l]} = w^{[l]}a^{[l-1]} + b^{[l]}$

$a^{[l]} = g^{[l]}(z^{[l]})$

Vectorizing forward propagation

$A^{[0]} = X$ input layer

$Z^{[l]} = W^{[l]}A^{[l-1]} + b^{[l]}$