Replicate Neurons in an Artificial Model

Artificial neuron modeling

One way to replicate nature’s network of neurons in an artificial model is to have layers of neurons, with each one connected to every other neuron in the preceding and subsequent layers. The following diagram illustrates this idea.

We can see the three layers, each with three artificial neurons, or nodes. We can also see how each node is connected to every other node in the preceding and next layer.

But what part of this cool looking architecture does the learning? What do we adjust in response to training examples? Is there a parameter that we can refine, like the slope of the linear classifiers we looked at earlier?

A neural network with weights

The most obvious thing is to adjust the strength of the connections between nodes. Within a node, we can adjust the summation of the inputs or we can adjust the shape of the sigmoid threshold function, but that’s more complicated than simply adjusting the strength of the connections between the nodes.

If the simpler approach works, let’s stick with it! The ...