Composition applies one function to another function’s output. In a multilayer neural network, each layer’s output becomes the next layer’s input. A weight matrix handles the whole layer at once, while the input is a vector. section 6.1.1
x_1 = f(w_1 * x + b_1) #1 x_2 = f(w_2 * x_1 + b_2) #2 ... o = f(w_h * x_n + b_h) #3
Follow the sequence of layer calculations: each result feeds the next layer, until the network produces its output. section 6.1.1
