We saw in the previous lecture that perceptrons have limited scope in the type of concepts they can learn - they can only learn linearly separable functions. However, we can think of constructing ...