Can an Algorithm Recognize Creativity?

One team at Yahoo Labs is using micro-videos to develop an algorithm for the identification of creativity.

|
Nov 25 2014, 7:00pm

Still from a Vine by Marlo Meekins

If creativity could be quantified, could it lead to better videos, stronger images, and richer content? One team at Yahoo Labs is using the Vine platform as a jumping-off point to develop a machine learning algorithm for the identification and definition of creativity. 

As reported in MIT's Technology Review, Yahoo Labs recently produced an eight-page paper on a study of 3,800 Vines that used analytics, machine learning, and 300 crowdsourced volunteers to create an algorithm with the ability to decide which short-form videos are creative and which are not. Spearheaded by Miriam Redi, the team first showed the few videos to humans tasked with deciding whether they personally felt that the Vines were creative. After that, the researchers used algorithms to analyze the content of the Vines, comparing these videos' applications of the rule of thirds and depth of field, as well as their uses of stop-motion and animation techniques, and whether or not they were designed to function as continuous loops. 

A graph showing the most relevant features and components for the prediction of creative videos. via

This research suggested that content itself is the most important factor in deciding whether or not videos are creative. Using this knowledge, Redi and her team were able to create an algorithm with the ability to "correctly classify videos as either creative or noncreative 80 per cent of the time," from a set of Vines it had never encountered before. The study concludes with the fact that it's totally possible for a machine learning algorithm to "automatically detect creative video, with promising results," provided that it can take both "aesthetic value and novelty" into consideration.

Basically, there's now an algorithm that can decide if your Vines are creative or not, and its potential applications range from the automatic curation of rich creative content to the possibility of machines one day being able to their own creative hyper-short films. 

Click here to check out 6 Seconds of Sound and Vision: Creativity in Micro-Videos, Miriam Redi and company's original study on machines learning and creativity on Vine. 

h/t Technology Review. Thumbnail image via

Related:

Video Artist Creates Dreamy Worlds Using Vine

This Awesome Claymation Throwback Was Made From A Compilation Of Vines

Anne Horel's "GIF & Vine" Series Explores Possibilities Of Hybrid A/V Looping

More VICE
Vice Channels