SoylentNews is people
SoylentNews
A machine learning algorithm has created tiny (64×64 pixels) 32-frame videos based on text descriptions:
The researchers trained the algorithm on 10 types of scenes, including "playing golf on grass," and "kitesurfing on the sea," which it then roughly reproduced. Picture grainy VHS footage. Nevertheless, a simple classification algorithm correctly guessed the intended action among six choices about half the time. (Sailing and kitesurfing were often mistaken for each other.) What's more, the network could also generate videos for nonsensical actions, such as "sailing on snow," and "playing golf at swimming pool," the team reported this month at a meeting of the Association for the Advancement of Artificial Intelligence in New Orleans, Louisiana.
[...] Currently, the videos are only 32 frames long—lasting about 1 second—and the size of a U.S. postage stamp, 64 by 64 pixels. Anything larger reduces accuracy, says Yitong Li, a computer scientist at Duke University in Durham, North Carolina, and the paper's first author. Because people often appear as distorted figures, a next step, he says, is using human skeletal models to improve movement.
Tuytelaars also sees applications beyond Hollywood. Video generation could lead to better compression if a movie can be stored as nothing but a brief description. It could also generate training data for other machine learning algorithms. For example, realistic video clips might help autonomous cars prepare for dangerous situations they would not frequently encounter. And programs that deeply understand the visual world could spin off useful applications in everything from refereeing to surveillance. They could help a self-driving car predict where a motorbike will go, for example, or train a household robot to open a fridge, Pirsiavash says.
An AI-generated Hollywood blockbuster may still be beyond the horizon, but in the meantime, we finally know what "kitesurfing on grass" looks like.
(Score: 2) by VLM on Sunday February 25 2018, @02:57PM
Its generally believed that pr0n leads technology, but there is a side path where about a decade ago the technology to turn free text into rap music famously launched the career of moonman on 4chan /pol/ (Um, don't google that at work NSFW).
Likewise I would suspect the most creative users of this technology will be 4chan /pol/ asking the video oracle for stuff like "gimmie birth of a nation but islamic themed" or "how about 1970s blaxploitation recast with soyboy hipster theme". At college we had to watch "Uptown Saturday Night", and I don't remember why, possibly as lame of a reason as the professor thought it was funny, that was a tolerable comedy in itself, but imagine that recast with the funniest stereotypes of modern hipsters. Actually, that might sell pretty well as a formulaic teen movie.
The other thing you'll see a heck of a lot of is "Modify the Zapruder film such that the secret service agent clearly shoots JFK instead of blurrily shoots him on the original" or whatever. Honestly though I think there was a Cuban hit team on the grassy knoll and to avoid global thermonuclear war in response, clear heads prevailed and the whole thing was covered up. It fits a lot of peculiarities in Cuban/USA (and Cuban/USSR) relations since 1960 or so. I admit it suffers a little from the rationalization argument that it makes far too much sense and matches evidence too well such that its a little unrealistic, I know from personal experience that real world military operations (admittedly not hard core stuff like political leader assassinations) are always a little screwed up in the details and nothing ever goes precisely according to plan.