Why is it that a perceptron or a single layered neural network can't solve the XOR problem, or problems that are linearly inseparable.
Before someone tells me to use google, I have. All the neural network guides that are online or in books are so overly complicated with masses of equations that describe everything - and the english language is totally abandoned. I can't get my head around why exactly it is that we cannot draw a line between the two and why do we even need to draw a line?!
The perceptron learning algorithm is something else I'm a bit confused on so help would be welcomed!