A good summary of what an artificial neural network is, but I think we need to emphasize much more strongly what it is not. It most definitely is not composed of neurons which are biological cells that exist only in living beings with nervous systems. It may or may not share any features, with (still hypothetical, possibly theoretical) neural networks that are said to be important structural and/or functional components of human (and some non human animals) brains and other parts of the nervous system. And, we must bear in mind that this hypothesis (or theory depending on your view of the strength of the data in support of it) of neural networks is not the only theory of brain structure/function, though it is the one most people are familiar with. In fact the only thing artificial neural networks for certain share in common with putative neural networks in the brain/body is the name.
Some argue that the features of the one (artificial neural networks) mirror the hypothesized features of the other (biological neural networks). It is fine to argue such a thing, however because the features of biological neural networks are only hypothetical they may not actually be features at all. They may not even exist at all. The so called “features” of biological neural networks may in fact be completely at odds with the reality of what is actually the case about the brain and nervous system. In contrast the features of artificial neural networks are totally known and describable as ably demonstrated in the very article upon which I am commenting.
I harp on this point because the conflating of artificial neural networks with (putative) biological neural networks is but one path down the slippery slope that leads to the compulogical fallacy. One we begin suggesting (through our words/terminology) that a non-biological/machine/computer thing like artificial neural networks are the same as or even very similar to supposed biological neural networks it becomes all that much easier to assign other characteristics/behaviors/attributes to machines that can only (logically only) be attributed/applied to human beings and some non human animals. This is how we have arrived where we are today with artificial “intelligence” and machines that “learn” when no such intelligent machines exist and machines cannot learn, machine learning being a term composed of two words that when combined in that order result in a logical contradiction and a thing which is logically impossible. To put it simply if a machine could learn it would no longer be a machine.