15.8 Concepts of neural networks

Now let’s generalize what we just saw. We keep the three input nodes and call this the input layer. These feed into the hidden layer, and the results are in the output layer.

Figure 15.15: Neural network with 3 inputs, 4 hidden nodes, and 2 outputs

Processing in a neural network moves from left to right, and we call this the “forward direction.”

Figure 15.15 shows the three input nodes, four nodes in the hidden layer, and two output nodes. I’ve also shown the weights w_i,1 for the input nodes going to the first hidden node h₁. We also have weights relating the input nodes to h₂, h₃, and h₄, and relating those to the output nodes. I have not shown them to keep the diagram less cluttered.

Another name for a node is neuron, and Figure 15.15 shows a neural network. More precisely, this is ...