(Image: Google)

Welcome to a psychedelic universe where aqueducts surrounded by glaciers fold in on themselves. This kaleidoscopic image, created by a quirk of brain-mimicking software, started out as something much less spectacular: random noise.

The software is an artificial neural network developed by Google, called the Inception module. Just as the brain detects patterns to classify images and make sense of speech, so neural networks can simulate this ability to identify features in an image.

(Image: Google)

But like children seeing shapes when looking at clouds, the software can over-interpret what it sees. This can lead to escalating absurdities, like the red landscape above which the network processed to include tractor-like vehicles and a dog emerging from a tree. In the pictures below, hybrid pig-snails appear in the sky.

(Image: Google)

(Image: Google)

Although the surreal landscapes produced may look pretty, they aren’t exactly useful. But by studying how the software processes information, Google’s engineers can tweak it to optimise its interpretative skills.

Neural networks’ talents aren’t limited to making sense of images. They can quickly estimate the number of objects in a scene without counting them – mimicking the ability of humans and some animals – and even play video games.

If you would like to reuse any content from New Scientist, either in print or online, please contact the syndication department first for permission. New Scientist does not own rights to photos, but there are a variety of licensing options available for use of articles and graphics we own the copyright to.

This entry passed through the Full-Text RSS service – if this is your content and you’re reading it on someone else’s site, please read the FAQ at fivefilters.org/content-only/faq.php#publishers.

After Click – secret code will be here:

Related Posts

Facebook Comments

Return to Top ▲Return to Top ▲