Advertisement

Researchers teach AI system to run complex physics experiment

"It did things a person wouldn't guess," said physicist Paul Wigley.

By Brooks Hays
Researchers trained an artificial intelligence system to manage a gas-trapping laser experiment. Photo by Stuart Hay/ANU
Researchers trained an artificial intelligence system to manage a gas-trapping laser experiment. Photo by Stuart Hay/ANU

ACTON, Australia, May 16 (UPI) -- A pair of physicists in Australia have trained an artificial intelligence system to replicate the experiment that won the 2001 Nobel Prize.

The experiment involves what is known as a Bose-Einstein condensate, the trapping of an ultra-cool gas in a series of lasers.

Advertisement

At just a billionth of a degree above absolute zero, Bose-Einstein condensates constitute some of the coldest temperatures in the universe -- colder than interstellar space.

Because these trapped gases are so sensitive, researchers are working to harness and utilize them as sensors for mineral exploration and navigational systems. A sensor powered by Bose-Einstein condensates could potentially pick up on tiny changes in the Earth's magnetic field or gravity.

Physicists Paul Wigley and Michael Hush cooled down gas to one microkelvin and ceded control on the laser system to their artificial intelligence system -- a system they trained to manage the delicate gas-trapping laser device.

The system then cooled the gas to a single nanokelvin using a variety of laser-manipulation techniques the researchers say humans likely never would have thought of.

"It did things a person wouldn't guess, such as changing one laser's power up and down, and compensating with another," Wigley, a researcher at the Australian National University, said in a news release.

Advertisement

The benefit of an AI-operated experiment is that it -- if incorporated into a scientific sensor -- could manage and calibrate the gas-trapping lasers much more efficiently.

"You could make a working device to measure gravity that you could take in the back of a car, and the artificial intelligence would recalibrate and fix itself no matter what," said Hush, a physicist at the University of New South Wales.

The researchers detailed their experiment in the journal Scientific Reports.

Latest Headlines