Artificial Intelligence By Example

Linear separability was achieved

Bear in mind that the whole purpose of this feedforward network with backpropagation through a cost function was to transform a linear non-separable function into a linearly separable function to implement classification of features presented to the system. In this case, the features had 0 or 1 value.

One of the core goals of a layer in a neural network is to make the input make sense, meaning to be able to separate one kind of information from another.

h1 and h2 will produce the Cartesian coordinate linear separability training axis, as implemented in the following code:

h1= h1 * -b1 h2= h2 * b2 print(h1,h2)

Running the program provides a view of the nonlinear input values once they have been trained by the ...