Question $6$

Having multiple perceptrons can actually solve the XOR problem satisfactorily: this is because each perceptron can partition off a linear part of the space itself, and they can then combine their results.

true $-$ perceptrons can do this but are unable to learn to do it $-$ they have to be explicitly hand$-$coded

true $-$ this works always, and these multiple perceptrons learn to classify even complex problems.

false $-$ perceptrons are mathematically incapable of solving linearly inseparable functions, no matter what you do

false $-$ just having a single perceptron is enough