|
Condiv posted:All of them came to their own conclusions about what art is and pursues it, as opposed to a neural net which can only reflect the biases of its creators. Actually, most artists don't know what the gently caress they are doing or why. Also, a neural network doesn't necessarily reflect the biases of its creators either. A neural network is trained using a method which you might imagine as similar to Pavlovian conditioning. The neural net is exposed to a set of training data, and the network's topology is reinforced if it produces a desirable answer, and mutated if it produces an undesirable answer. Eventually the network is arranged in a way that transforms the input into the corresponding desirable output. It is an uninteresting (in my opinion) numerical technique which results in a system which will happen to be correct within a given error value. It has more in common with something like a linear regression than human intelligence, despite its name.
|
#
¿
Nov 29, 2016 13:09
|
|
|
|
| # ¿ May 15, 2024 13:23 |
|
|
Condiv posted:Yes I'm aware of how neural networks work, I've created one of my own. The desirable/undesirable answer is where the creators biases are introduced and why the network ends up reflecting those who train it. Ah, I think I just misunderstood you. My thinking was that the definition of what constitutes a desirable or undesirable answer is the entire motivation for creating the network, and not a mere reflection of bias. Regarding your second line, I wasn't arguing that artists are not cognizant in general. I do believe artists possess cognition (nearly all living humans do). I was saying that many have not come to any conclusions about what art is and how best to pursue their artistic endeavours. Sort of beside the point, perhaps.
|
|
#
¿
Nov 29, 2016 13:36
|
|
|
Condiv posted:The point is that although there is derivation in art, there is a lot more to it that involves the artists desires, life experiences, and personality which highly specialized neural nets cannot reproduce. Right now neural nets are only really good at creating derivative work based on the biases of its trainers. I think its irrelevant anyway because art objects are independent of their creator, and the circumstances of their creation. You don't need a neural network or anything to create art objects, you could just use a random number generator. If a computer produces an art object, we must judge the merit of the art object on its own terms, not in light of the fact that it was produced by a computer. Recently a computer produced a new painting in the style of Rembrandt. I don't know poo poo about Rembrandt or painting in general for that matter, so I have no idea if its a 'good' painting or not, but the point is you can criticize it as a painting in the same way that you would criticize Rembrandt's actual paintings. It doesn't matter that a machine made it. Art objects are not people, and art objects are independent of their creators.
|
|
#
¿
Nov 29, 2016 20:16
|
|
|
A Wizard of Goatse posted:The whole discursion about authorial intent and 'meaning' is pointless; computer-generated art is bland on its own merits not because computers lack qualia deep within their circuits but because they don't innovate, they'll just repeat minor variations on the same basic task endlessly based on the parameters humans feed them. I think its interesting that randomly generating art could possibly produce innovations that we can understand, whereas more sophisticated techniques would be limited in the ways you describe.
|
|
#
¿
Nov 29, 2016 20:24
|
|
|
A Wizard of Goatse posted:A million monkeys banging on a million typewriters for a million years might eventually produce the complete works of Shakespeare, but it's gonna take a lot more man-hours to find the Shakespeare in all the gibberish than it did for Shakespeare to just write it. You're right that it is not an efficient method of generating quality art objects, but also I think optimality is irrelevant here.
|
|
#
¿
Nov 29, 2016 20:28
|
|
|
Condiv posted:no, but my argument has never been that computers can't generate good or entertaining things. they just can't generate anything with meaning without a human behind them. If a computer generated the text of Albert Camus' "The Stranger", why is that less "meaningful" than Camus writing the text himself.
|
|
#
¿
Nov 29, 2016 20:30
|
|
|
A Wizard of Goatse posted:How so? If you're not concerned with reducing the conscious human labor involved, why even bother with computers? You can poke around on the ground until you find a pretty rock. The "random art generating computer" was just a thought experiment to illustrate that computers are capable of producing unique, innovative art. I am not saying this is a worthwhile endeavor in practice.
|
|
#
¿
Nov 29, 2016 20:41
|
|
|
Condiv posted:it would have no meaning behind it because meaning requires cognition. i may find meaning in it, but as i pointed out earlier, there's a difference from finding meaning in something and something having meaning. I think the only way to reconcile our disagreement at this point would be to begin discussing what you mean by 'meaning', the nature of art, art objects, and beauty. But we should not derail this thread and turn it into a discussion about aesthetics.
|
|
#
¿
Nov 29, 2016 20:42
|
|
|
|
| # ¿ May 15, 2024 13:23 |
|
|
A Wizard of Goatse posted:Yeah but what makes you say it's the computer producing the art, rather than whoever fishes the unique, innovative-looking stuff from the millions upon millions of copies of basically the same lovely thing? Without that guy, you're never going to get art anyone wants to look at; the random generator is not capable of producing good art on its own it is simply the medium that guy works in. You could hypothetically have a standalone art machine that produced interesting work, and there's no reason it would need a genuine sense of self to do so, but it'd have to be a very different sort of thing than modern computer programs are. Ah I see what you're saying. That's interesting. Art as a search problem. So if we imagine that art works are not generated or created, but rather discovered, the task of the searcher in our thought experiment is identical to the task of the artist. I'll have to think about that!
|
|
#
¿
Nov 29, 2016 20:54
|
|