How physics can improve image-generating AI
The laws governing electromagnetism and even the weak nuclear force could be worth mimicking
TODAY’S BEST image-generating artificial-intelligence (AI) models are remarkable. Ask OpenAI’s DALL-E 3, or its counterparts Midjourney and Stable Diffusion, to draw a penguin sipping on a vodka martini on the French Riviera and they will do so with aplomb. Ask them to replicate it in the style of Rembrandt or Caravaggio and they will speedily oblige.
These abilities all depend on a family of physics-inspired algorithms known as diffusion models. For now, they reign supreme. But that may not always be the case. A team of physicists and computer scientists at the Massachusetts Institute of Technology (MIT) has been taking inspiration from the laws of nature to come up with a series of increasingly sophisticated algorithms that can generate higher-quality images faster, and with smaller training data sets, than diffusion models.
This article appeared in the Science & technology section of the print edition under the headline “Painting by numbers”
A flower’s female sex organs can speed up fertilisation
They can also stop it from happening
The dominant model of the universe is creaking
Dark energy could break it apart
Only 5% of therapies tested on animals are approved for human use
More rigorous experiments could improve those odds
A flower’s female sex organs can speed up fertilisation
They can also stop it from happening
The dominant model of the universe is creaking
Dark energy could break it apart
Only 5% of therapies tested on animals are approved for human use
More rigorous experiments could improve those odds
The secret to taking better penalties
Practise with an augmented-reality headset
China has become a scientific superpower
From plant biology to superconductor physics the country is at the cutting edge
Like people, elephants call each other by name
And anthropoexceptionalism takes another tumble