Space and Physics
PUBLISHED
Artificial intelligence might not be smarter than us yet, but it is getting surprisingly good – even in fields you don't expect. The latest one is a physicist AI machine that was put in charge of creating a rare state of matter, and it did it better than human scientists.
The rest of this article is behind a paywall. Please sign in or subscribe to access the full content.The AI was designed to understand and improve an experiment to form a Bose-Einstein condensate, a special type of matter obtained when certain types of gas are cooled to almost absolute zero.
To achieve the state, very cold gas is cooled through laser cooling, a technique that uses multiple lasers to take away energy from single atoms and thus reduce their temperature. This type of experiment won physicists Eric Cornell, Wolfgang Ketterle, and Carl Wieman the Nobel Prize in Physics in 2001.
In a paper published by Scientific Reports, researchers from the Australian National University explained their new approach to problem solving. The AI machine learning algorithm was asked to learn how to create Bose-Einstein condensate, and it did so in record time.
"I didn't expect the machine could learn to do the experiment itself, from scratch, in under an hour," said co-lead researcher Paul Wigley in a statement. "A simple computer program would have taken longer than the age of the universe to run through all the combinations and work this out."
The researchers cooled gas to around 1 microkelvin, and then the AI took over. Controlling three laser beams, the AI cooled the gas down to 1 nanokelvin. The AI's approach was something the researchers had never seen before.
"It did things a person wouldn't guess, such as changing one laser's power up and down, and compensating with another," said Mr. Wigley. "It may be able to come up with complicated ways humans haven't thought of to get experiments colder and make measurements more precise."
Although this AI was designed with Bose-Einstein condensate in mind, the researchers think that machine learning can be applied to many experimental setups, and one day soon AIs might become a fundamental instrument in the physicists’ tool belt.
