Artificial Intelligence By Example

A simplified version of a cost function and gradient descent

Now the system must train. To do that, we need to measure the number of predictions, 1 to 4, that are correct at each iteration and decide how to change the weights/biases until we obtain proper results.

A slightly more complex gradient descent will be described in the next chapter. In this chapter, only a one-line equation will do the job. The only thing to bear in mind as an unconventional thinker is: so what? The concept of gradient descent is minimizing loss or errors between the present result and a goal to attain.

First, a cost function is needed.

There are four predicates (0-0, 1-1, 1-0, 0-1) to train correctly. We simply need to find out how many are correctly trained at ...