Oh, I even love the word itself, but what are they, really? Well, it is a type of artificial neuron that takes several binary inputs and produces a single binary output.

What one needs to know are the binary inputs and their relative weight in the decision-making process (…) by varying the weights and the threshold, we can get different models of decision-making (…) a perceptron can weigh up different kinds of evidence in order to make decisions (…) another way perceptrons can be used is to compute the elementary logical functions we usually think of as underlying computation, functions such as AND, OR, and NAND.

References

Michael Nielsen, Neural Networks and Deep Learning, Chapter 01, full book available here