5 ms·
Some of these images and those from similar projects could be in an art gallery. They are art; provoking original, emotional, responses. Most people hear abou
by WhitneyLand 10y ago
Some of these images and those from similar projects could be in an art gallery. They are art; provoking original, emotional, responses.
Most people hear about self-driving cars, but not about the fact that machines have already begun to emulate human creativity in the most intimate way. For a while, this secret assault on our uniqueness will stay among us.
- yolesaber 10y agoHe built an H.R. Giger [0] generator, essentially. [0] https://en.wikipedia.org/wiki/Penis_Landscape https://en.wikipedia.org/wiki/Penis_Landscape
- TheSpiceIsLife 10y agoI thought the images looked mildly interesting but lack a certain sense of Giger. With a bunch of Giger images plus a bunch of porn as input the output would probably be fairly indistinguishable from Giger originals.
- PaulAJ 10y agoMy immediate thought was that they looked like the work of Salvador Dalí
- qbrass 10y agoSounds like the premise for the strangest Viagra ad ever.
- eveningcoffee 10y agoCalm your horses. Generative nonsense have been already explored in literature in 90s. So for me it does not produce any new feeling. Now for a newer generation playing with a real ding is perhaps a little more intriguing.
- WhitneyLand 10y agoActually, no. Nothing produced in the 90's is the same as what's been done in the last few years. But please cite your best work and we'll have a look. Regarding your dick joke, it's not really relevant. This is not interesting because it is pornographic.
- eveningcoffee 10y agoThe work of 90s as I know it (sorry, no references, I was quite young then, do not care to loop up at the moment), was taking some corpus (and not an English one) and letting a generative algorithm run on it. This was not a research but art, literature, and as such, actually could have been very regional. Of course the current state of art is on the different level but in basic principles not very different and as such not very surprising as an form of art. PS. it does not imply that no interesting results may grow out of this.
- pmyjavec 10y agoI disagree, because the machine is still using human generated input for creating these images. Not to take away from what's been created here because it's interesting to look at, but an image filter could be considered an attack on our uniqueness because it takes an existing human art work and can potentially create something more interesting, but most people wouldn't consider it an assault. I get this is a "neural net", which makes people think these paintings are created by the terminator, but these images still have very real human inputs which is why we can relate to them.
- Houshalter 10y agoWhat human input? I mean the pictures were taken by humans. But they weren't drawn by humans, they weren't meant to be art, and no one intended the output of the network to look the way it does. There is very little human input here.
- gridspy 10y agoIf I pick a shell I find beautiful from the shore, did the beach create the shell, or did I? Computer algorithms produce a lot of trash. Just because the human didn't wield a brush, as long as we're seeing filtered and tuned results, a human had a hand in creating these.
- WhitneyLand 10y agoYes humans were instrumental, but we can acknowledge that while still accepting their role as separate from artist: http://www.nytimes.com/1996/06/30/movies/how-creative-artists-court-the-muse.html http://www.nytimes.com/1996/06/30/movies/how-creative-artist...
- santaclaus 10y ago> already begun to emulate human creativity in the most intimate way I'd argue that the creative entity here is @gabeeegoooh. The machine didn't make the (provocative) decision to synthesize NSFW images - that artistic stroke was painted by the fingers and mind that coded up the adversarial neural net.
- WhitneyLand 10y agoOne way to look at @gabeeegoooh's role is as muse; He provided the "inspiration" for the creativity.
- wpietri 10y agoDefinitely. If creativity were merely the creation of interesting images, we'd have to call nature far more creative than any human artist. The Grand Canyon is beautiful, but we don't put up a plaque giving the Colorado River credit for its thoughtful use of form and color. Calling the neural network creative is like crediting Van Gogh's brushes for Van Gogh's brushwork. It's a possible use for the word, but it's very different use than the one normally meant.
- WhitneyLand 10y agoVan Gogh's brushes were simple tools used to render a direct expression of his intent. These networks are more akin to a child. You teach her to paint, give her input and feedback. The result may be something emotionally moving that you could have not even conceived of. Your opinion and mine are both surely subjective. However, it seems it will only become more difficult to split the hairs required to argue that machine learning creativity is fundamentally different from our own.
- wpietri 10y agoBrushes are not pure transmitters of intent the way that, say, Photoshop is. Brushes (and paints and most artistic tools) have an artistic character that makes its own contribution to the artwork. I also think you ignore the interplay between medium and artist, the way intent evolves both during the course of an artist's careeer and during the production of a single work. I don't think I'm hair-splitting here. Creativity is something conscious entities do. Once we have conscious machines, they could well be creative. But until then, the creativity is the artists, even if they are using generative systems to create art. And that's nothing new. Every landscape painter and photographer is using a generative system, nature, to find interesting things worth turning into art. Giving the neural networks credit is akin to crediting the haystacks rather than Monet. More explicitly generative art, where humans set up systems to produce works, goes back hundreds of years. That interesting patterns can emerge out of randomness plus systems is obvious to anybody who has played Yahtzee. That we use more dice and more complicated systems does not change where authorship lies. Bigger systems may produce more novel outputs, but novelty and creativity aren't the same thing.