An AI Dreamed Up Street Scenes, and They’re Surprisingly Good
You're looking at pure fiction: this image was actually created by an AI, trained on the kinds of driver's-eye labeled images often supplied to self-driving cars. Usually, humans describe which parts of a picture are, say, cars or sidewalks, and the labeled images are used to train neural networks to recognize what they're looking at. Instead, Qifeng Chen, from Stanford University and Intel, got a similar neural net to use those learnings to render new street scenes. It puts a road somewhere down the middle, trees down the side, cars on the road ... and the results are surprisingly good.
Chen tells New Scientist that the software could be used to great effect in video games, where it could create realistic virtual worlds on the fly. It's worth bearing in mind that the games industry is big business: Twitter famously invested $150 million in Magic Pony, which also uses AI to generate high-quality computer-game graphics, giving the startup a valuation of a cool $1 billion.
Keep Reading
Most Popular
Large language models can do jaw-dropping things. But nobody knows exactly why.
And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.
How scientists traced a mysterious covid case back to six toilets
When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.
The problem with plug-in hybrids? Their drivers.
Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.